{ "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, "head_dim": 192, "hidden_act": "silu", "hidden_size": 6144, "index_head_dim": 128, "index_n_heads": 32, "index_share_for_mtp_iteration": true, "index_skip_topk_offset": 3, "index_topk": 2048, "index_topk_freq": 4, "index_topk_pattern": null, "indexer_rope_interleave": true, "indexer_types": [ "full", "full", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared", "full", "shared", "shared", "shared" ], "initializer_range": 0.02, "intermediate_size": 12288, "kv_lora_rank": 512, "max_position_embeddings": 1048576, "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, "quantization_config": { "autoround_version": "0.14.0", "batch_size": 2, "bits": 4, "block_name_to_quantize": [ "^model\\.layers\\.3$", "^model\\.layers\\.4$", "^model\\.layers\\.5$", "^model\\.layers\\.6$", "^model\\.layers\\.7$", "^model\\.layers\\.8$", "^model\\.layers\\.9$", "^model\\.layers\\.10$", "^model\\.layers\\.11$", "^model\\.layers\\.12$", "^model\\.layers\\.13$", "^model\\.layers\\.14$", "^model\\.layers\\.15$", "^model\\.layers\\.16$", "^model\\.layers\\.17$", "^model\\.layers\\.18$", "^model\\.layers\\.19$", "^model\\.layers\\.20$", "^model\\.layers\\.21$", "^model\\.layers\\.22$", "^model\\.layers\\.23$", "^model\\.layers\\.24$", "^model\\.layers\\.25$", "^model\\.layers\\.26$", "^model\\.layers\\.27$", "^model\\.layers\\.28$", "^model\\.layers\\.29$", "^model\\.layers\\.30$", "^model\\.layers\\.31$", "^model\\.layers\\.32$", "^model\\.layers\\.33$", "^model\\.layers\\.34$", "^model\\.layers\\.35$", "^model\\.layers\\.36$", "^model\\.layers\\.37$", "^model\\.layers\\.38$", "^model\\.layers\\.39$", "^model\\.layers\\.40$", "^model\\.layers\\.41$", "^model\\.layers\\.42$", "^model\\.layers\\.43$", "^model\\.layers\\.44$", "^model\\.layers\\.45$", "^model\\.layers\\.46$", "^model\\.layers\\.47$", "^model\\.layers\\.48$", "^model\\.layers\\.49$", "^model\\.layers\\.50$", "^model\\.layers\\.51$", "^model\\.layers\\.52$", "^model\\.layers\\.53$", "^model\\.layers\\.54$", "^model\\.layers\\.55$", "^model\\.layers\\.56$", "^model\\.layers\\.57$", "^model\\.layers\\.58$", "^model\\.layers\\.59$", "^model\\.layers\\.60$", "^model\\.layers\\.61$", "^model\\.layers\\.62$", "^model\\.layers\\.63$", "^model\\.layers\\.64$", "^model\\.layers\\.65$", "^model\\.layers\\.66$", "^model\\.layers\\.67$", "^model\\.layers\\.68$", "^model\\.layers\\.69$", "^model\\.layers\\.70$", "^model\\.layers\\.71$", "^model\\.layers\\.72$", "^model\\.layers\\.73$", "^model\\.layers\\.74$", "^model\\.layers\\.75$", "^model\\.layers\\.76$", "^model\\.layers\\.77$", "model.layers" ], "data_type": "int", "extra_config": { ".*eh_proj.*": { "bits": 16, "data_type": "float" }, ".*layers\\.0\\..*": { "bits": 16, "data_type": "float" }, ".*layers\\.1\\..*": { "bits": 16, "data_type": "float" }, ".*layers\\.2\\..*": { "bits": 16, "data_type": "float" }, ".*layers\\.[0-2]\\.mlp.*": { "bits": 16, "data_type": "float" }, ".*layers\\\\\\.[0-2]\\\\\\..*": { "bits": 16, "data_type": "float" }, ".*self_attn.*": { "bits": 16, "data_type": "float" }, ".*self_attn\\.indexer.*": { "bits": 16, "data_type": "float" }, ".*shared_experts.*": { "bits": 16, "data_type": "float" }, ".*weights_proj.*": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.0.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.1.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.10.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.11.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.12.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.13.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.14.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.15.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.16.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.17.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.18.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.19.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.2.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.20.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.21.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.22.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.23.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.24.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.25.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.26.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.27.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.28.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.29.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.3.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.30.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.31.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.32.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.33.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.34.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.35.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.36.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.37.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.38.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.39.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.4.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.40.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.41.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.42.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.43.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.44.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.45.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.46.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.47.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.48.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.49.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.5.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.50.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.51.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.52.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.53.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.54.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.55.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.56.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.57.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.58.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.59.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.6.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.60.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.61.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.62.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.63.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.64.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.65.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.66.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.67.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.68.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.69.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.7.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.70.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.71.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.72.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.73.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.74.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.75.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.76.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.77.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.8.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.indexer.wk": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.indexer.wq_b": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.kv_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.o_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.q_a_proj": { "bits": 16, "data_type": "float" }, "model.layers.9.self_attn.q_b_proj": { "bits": 16, "data_type": "float" }, "model.layers.78.eh_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.mlp.gate": { "bits": 16, "data_type": "fp" }, "model.layers.78.mlp.shared_experts.down_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.mlp.shared_experts.gate_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.mlp.shared_experts.up_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.indexer.weights_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.indexer.wk": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.indexer.wq_b": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.kv_a_proj_with_mqa": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.kv_b_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.o_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.q_a_proj": { "bits": 16, "data_type": "fp" }, "model.layers.78.self_attn.q_b_proj": { "bits": 16, "data_type": "fp" } }, "gradient_accumulate_steps": 4, "group_size": 64, "low_gpu_mem_usage": true, "nsamples": 512, "packing_format": "auto_round:auto_gptq", "quant_method": "auto-round", "sym": true }, "rms_norm_eps": 1e-05, "rope_interleave": true, "rope_parameters": { "rope_theta": 8000000, "rope_type": "default" }, "routed_scaling_factor": 2.5, "scoring_func": "sigmoid", "tie_word_embeddings": false, "topk_group": 1, "topk_method": "noaux_tc", "transformers_version": "5.10.2", "use_cache": true, "v_head_dim": 256, "vocab_size": 154880 }