| { |
| "architectures": [ |
| "HYV3ForCausalLM" |
| ], |
| "bos_token_id": 120000, |
| "enable_attention_fp32_softmax": false, |
| "enable_lm_head_fp32": true, |
| "enable_moe_fp32_combine": false, |
| "eod_token_id": 120026, |
| "eos_token_id": 120025, |
| "expert_hidden_dim": 1536, |
| "moe_intermediate_size": 1536, |
| "first_k_dense_replace": 1, |
| "head_dim": 128, |
| "hidden_act": "silu", |
| "hidden_size": 4096, |
| "initializer_range": 0.006, |
| "intermediate_size": 13312, |
| "max_position_embeddings": 262144, |
| "model_type": "hy_v3", |
| "moe_router_enable_expert_bias": true, |
| "moe_router_use_sigmoid": true, |
| "num_attention_heads": 64, |
| "num_experts": 192, |
| "num_experts_per_tok": 8, |
| "num_hidden_layers": 80, |
| "num_key_value_heads": 8, |
| "num_shared_experts": 1, |
| "output_router_logits": true, |
| "pad_token_id": 120002, |
| "qk_norm": true, |
| "rms_norm_eps": 1e-05, |
| "rope_parameters": { |
| "rope_theta": 11158840.0, |
| "rope_type": "default" |
| }, |
| "route_norm": true, |
| "router_scaling_factor": 2.826, |
| "sep_token_id": 120007, |
| "tie_word_embeddings": false, |
| "transformers_version": "5.6.0", |
| "use_cache": true, |
| "use_grouped_mm": false, |
| "vocab_size": 120832, |
| "num_nextn_predict_layers": 0, |
| "quantization": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine", |
| "model.embed_tokens": { |
| "bits": 6, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "lm_head": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.53.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.54.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.55.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.56.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.57.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.58.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.59.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.60.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.61.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.62.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.63.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.64.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.65.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.66.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.67.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.68.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.69.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.70.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.71.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.72.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.73.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.74.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.75.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.76.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.77.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.78.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.self_attn.q_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.self_attn.k_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.self_attn.v_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.self_attn.o_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.shared_mlp.gate_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.shared_mlp.up_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.shared_mlp.down_proj": { |
| "bits": 8, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.switch_mlp.gate_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.switch_mlp.up_proj": { |
| "bits": 2, |
| "group_size": 128, |
| "mode": "affine" |
| }, |
| "model.layers.79.mlp.switch_mlp.down_proj": { |
| "bits": 3, |
| "group_size": 128, |
| "mode": "affine" |
| } |
| }, |
| "_name_or_path": "Hy3-JANG_2K", |
| "runtime": { |
| "bundle_has_mtp": false, |
| "mtp_layers": 0, |
| "mtp_mode": "dropped_for_smallest_affine", |
| "mtp_num_speculative_tokens": 0, |
| "mtp_status": "MTP dropped (smallest-affine profile)" |
| }, |
| "capabilities": { |
| "reasoning_parser": "qwen3", |
| "tool_parser": "hunyuan", |
| "think_in_template": false, |
| "supports_tools": true, |
| "supports_thinking": true, |
| "family": "hy_v3", |
| "modality": "text", |
| "modalities": { |
| "text": true, |
| "vision": false, |
| "audio": false, |
| "video": false |
| }, |
| "has_vision": false, |
| "has_audio": false, |
| "has_video": false, |
| "cache_type": "kv" |
| } |
| } |
