| { |
| "_debug_force_load_balance": false, |
| "architectures": [ |
| "MotifForCausalLM" |
| ], |
| "attention_cls": "gdla", |
| "attention_dropout": 0.0, |
| "auto_map": { |
| "AutoConfig": "configuration_motif.MotifConfig", |
| "AutoModel": "modeling_motif.MotifForCausalLM", |
| "AutoModelForCausalLM": "modeling_motif.MotifForCausalLM" |
| }, |
| "diff_v2": true, |
| "dtype": "bfloat16", |
| "elementwise_attn_output_gate": true, |
| "eos_token_id": [ |
| 0, |
| 3, |
| 6 |
| ], |
| "experts_top_k": 8, |
| "head_dim": 192, |
| "headwise_attn_output_gate": false, |
| "hidden_act": "poly_norm", |
| "hidden_clamp": 1000000.0, |
| "hidden_size": 4096, |
| "initializer_range": 0.02, |
| "interleave_moe_layer_step": 1, |
| "intermediate_size": 12288, |
| "k_ratio": 1, |
| "kv_lora_rank": 512, |
| "load_balance_coeff": 0.0001, |
| "max_position_embeddings": 262144, |
| "max_window_layers": 9, |
| "mhc_enabled": true, |
| "mhc_expansion_rate": 4, |
| "mhc_identity_init": false, |
| "mhc_sinkhorn_iters": 20, |
| "model_type": "Motif", |
| "moe_intermediate_size": 1280, |
| "mscale": 1.0, |
| "n_dense_first_layers": 2, |
| "num_attention_heads": 80, |
| "num_experts": 384, |
| "num_hidden_layers": 53, |
| "num_key_value_heads": 16, |
| "num_nextn_predict_layers": 1, |
| "num_noise_heads": 16, |
| "num_shared_experts": 1, |
| "original_seq_len": 4096, |
| "output_router_logits": false, |
| "polynorm_bias_clamp": 0.5, |
| "polynorm_output_scale": 0.5, |
| "polynorm_output_scale_per_layer": {}, |
| "q_lora_rank": 1024, |
| "qk_rope_head_dim": 64, |
| "quantization": { |
| "group_size": 64, |
| "bits": 2, |
| "mode": "affine", |
| "model.embed_tokens": { |
| "group_size": 64, |
| "bits": 6, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "lm_head": { |
| "group_size": 64, |
| "bits": 6, |
| "mode": "affine" |
| } |
| }, |
| "quantization_config": { |
| "group_size": 64, |
| "bits": 2, |
| "mode": "affine", |
| "model.embed_tokens": { |
| "group_size": 64, |
| "bits": 6, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.0.mlp.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.1.mlp.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.2.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.3.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.4.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.5.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.6.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.7.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.8.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.9.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.10.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.11.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.12.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.13.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.14.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.15.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.16.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.17.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.18.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.19.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.20.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.21.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.22.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.23.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.24.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.25.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.26.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.27.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.28.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.29.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.30.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.31.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.32.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.33.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.34.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.35.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.36.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.37.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.38.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.39.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.40.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.41.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.42.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.43.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.44.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.45.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.46.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.47.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.48.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.49.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.50.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.51.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_a": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wq_b_gate": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wkv_a": { |
| "group_size": 64, |
| "bits": 8, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wkv_b": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.self_attn.wo": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.gate_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.up_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.switch_mlp.down_proj": { |
| "group_size": 128, |
| "bits": 2, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.gate_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.up_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "model.layers.52.mlp.shared_experts.down_proj": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "lm_head": { |
| "group_size": 64, |
| "bits": 6, |
| "mode": "affine" |
| } |
| }, |
| "rms_norm_eps": 1e-05, |
| "rope_cos_sin_scale": 1.0, |
| "rope_factor": 64.0, |
| "rope_theta": 10000.0, |
| "route_norm": true, |
| "route_scale": 2.0, |
| "router_aux_loss_coef": 0.0, |
| "score_before_experts": false, |
| "score_func": "sigmoid", |
| "sliding_window": 128, |
| "sliding_window_pattern": "interleave", |
| "sliding_window_period": 4, |
| "swa_rope_theta": 10000.0, |
| "tie_word_embeddings": false, |
| "transformers_version": "5.7.0", |
| "use_cache": true, |
| "use_sliding_window": true, |
| "v_head_dim": 128, |
| "vocab_size": 220160 |
| } |
