{ "architectures": [ "Qwen3MoeForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 151643, "decoder_sparse_step": 1, "eos_token_id": 151645, "head_dim": 128, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "intermediate_size": 6144, "max_position_embeddings": 40960, "max_window_layers": 48, "mlp_only_layers": [], "model_type": "qwen3_moe", "moe_intermediate_size": 768, "norm_topk_prob": true, "num_attention_heads": 32, "num_experts": 128, "num_experts_per_tok": 8, "num_hidden_layers": 48, "num_key_value_heads": 4, "output_router_logits": false, "quantization": { "group_size": 64, "bits": 4, "model.embed_tokens": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.0.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.q_norm": false, "model.layers.0.self_attn.k_norm": false, "model.layers.0.self_attn.rope": false, "model.layers.0.input_layernorm": false, "model.layers.0.post_attention_layernorm": false, "model.layers.0.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.1.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.1.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.q_norm": false, "model.layers.1.self_attn.k_norm": false, "model.layers.1.self_attn.rope": false, "model.layers.1.input_layernorm": false, "model.layers.1.post_attention_layernorm": false, "model.layers.1.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.2.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.2.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.q_norm": false, "model.layers.2.self_attn.k_norm": false, "model.layers.2.self_attn.rope": false, "model.layers.2.input_layernorm": false, "model.layers.2.post_attention_layernorm": false, "model.layers.2.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.3.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.3.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.q_norm": false, "model.layers.3.self_attn.k_norm": false, "model.layers.3.self_attn.rope": false, "model.layers.3.input_layernorm": false, "model.layers.3.post_attention_layernorm": false, "model.layers.3.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.4.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.4.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.q_norm": false, "model.layers.4.self_attn.k_norm": false, "model.layers.4.self_attn.rope": false, "model.layers.4.input_layernorm": false, "model.layers.4.post_attention_layernorm": false, "model.layers.4.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.5.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.5.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.q_norm": false, "model.layers.5.self_attn.k_norm": false, "model.layers.5.self_attn.rope": false, "model.layers.5.input_layernorm": false, "model.layers.5.post_attention_layernorm": false, "model.layers.5.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.6.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.q_norm": false, "model.layers.6.self_attn.k_norm": false, "model.layers.6.self_attn.rope": false, "model.layers.6.input_layernorm": false, "model.layers.6.post_attention_layernorm": false, "model.layers.6.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.q_norm": false, "model.layers.7.self_attn.k_norm": false, "model.layers.7.self_attn.rope": false, "model.layers.7.input_layernorm": false, "model.layers.7.post_attention_layernorm": false, "model.layers.7.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.8.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.q_norm": false, "model.layers.8.self_attn.k_norm": false, "model.layers.8.self_attn.rope": false, "model.layers.8.input_layernorm": false, "model.layers.8.post_attention_layernorm": false, "model.layers.8.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.9.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.q_norm": false, "model.layers.9.self_attn.k_norm": false, "model.layers.9.self_attn.rope": false, "model.layers.9.input_layernorm": false, "model.layers.9.post_attention_layernorm": false, "model.layers.9.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.q_norm": false, "model.layers.10.self_attn.k_norm": false, "model.layers.10.self_attn.rope": false, "model.layers.10.input_layernorm": false, "model.layers.10.post_attention_layernorm": false, "model.layers.10.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.11.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.q_norm": false, "model.layers.11.self_attn.k_norm": false, "model.layers.11.self_attn.rope": false, "model.layers.11.input_layernorm": false, "model.layers.11.post_attention_layernorm": false, "model.layers.11.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.12.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.q_norm": false, "model.layers.12.self_attn.k_norm": false, "model.layers.12.self_attn.rope": false, "model.layers.12.input_layernorm": false, "model.layers.12.post_attention_layernorm": false, "model.layers.12.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.q_norm": false, "model.layers.13.self_attn.k_norm": false, "model.layers.13.self_attn.rope": false, "model.layers.13.input_layernorm": false, "model.layers.13.post_attention_layernorm": false, "model.layers.13.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.14.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.q_norm": false, "model.layers.14.self_attn.k_norm": false, "model.layers.14.self_attn.rope": false, "model.layers.14.input_layernorm": false, "model.layers.14.post_attention_layernorm": false, "model.layers.14.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.15.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.q_norm": false, "model.layers.15.self_attn.k_norm": false, "model.layers.15.self_attn.rope": false, "model.layers.15.input_layernorm": false, "model.layers.15.post_attention_layernorm": false, "model.layers.15.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.q_norm": false, "model.layers.16.self_attn.k_norm": false, "model.layers.16.self_attn.rope": false, "model.layers.16.input_layernorm": false, "model.layers.16.post_attention_layernorm": false, "model.layers.16.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.17.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.q_norm": false, "model.layers.17.self_attn.k_norm": false, "model.layers.17.self_attn.rope": false, "model.layers.17.input_layernorm": false, "model.layers.17.post_attention_layernorm": false, "model.layers.17.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.18.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.q_norm": false, "model.layers.18.self_attn.k_norm": false, "model.layers.18.self_attn.rope": false, "model.layers.18.input_layernorm": false, "model.layers.18.post_attention_layernorm": false, "model.layers.18.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.q_norm": false, "model.layers.19.self_attn.k_norm": false, "model.layers.19.self_attn.rope": false, "model.layers.19.input_layernorm": false, "model.layers.19.post_attention_layernorm": false, "model.layers.19.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.20.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.q_norm": false, "model.layers.20.self_attn.k_norm": false, "model.layers.20.self_attn.rope": false, "model.layers.20.input_layernorm": false, "model.layers.20.post_attention_layernorm": false, "model.layers.20.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.21.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.q_norm": false, "model.layers.21.self_attn.k_norm": false, "model.layers.21.self_attn.rope": false, "model.layers.21.input_layernorm": false, "model.layers.21.post_attention_layernorm": false, "model.layers.21.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.q_norm": false, "model.layers.22.self_attn.k_norm": false, "model.layers.22.self_attn.rope": false, "model.layers.22.input_layernorm": false, "model.layers.22.post_attention_layernorm": false, "model.layers.22.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.23.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.q_norm": false, "model.layers.23.self_attn.k_norm": false, "model.layers.23.self_attn.rope": false, "model.layers.23.input_layernorm": false, "model.layers.23.post_attention_layernorm": false, "model.layers.23.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.24.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.q_norm": false, "model.layers.24.self_attn.k_norm": false, "model.layers.24.self_attn.rope": false, "model.layers.24.input_layernorm": false, "model.layers.24.post_attention_layernorm": false, "model.layers.24.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.q_norm": false, "model.layers.25.self_attn.k_norm": false, "model.layers.25.self_attn.rope": false, "model.layers.25.input_layernorm": false, "model.layers.25.post_attention_layernorm": false, "model.layers.25.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.26.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.q_norm": false, "model.layers.26.self_attn.k_norm": false, "model.layers.26.self_attn.rope": false, "model.layers.26.input_layernorm": false, "model.layers.26.post_attention_layernorm": false, "model.layers.26.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.27.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.q_norm": false, "model.layers.27.self_attn.k_norm": false, "model.layers.27.self_attn.rope": false, "model.layers.27.input_layernorm": false, "model.layers.27.post_attention_layernorm": false, "model.layers.27.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.q_norm": false, "model.layers.28.self_attn.k_norm": false, "model.layers.28.self_attn.rope": false, "model.layers.28.input_layernorm": false, "model.layers.28.post_attention_layernorm": false, "model.layers.28.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.29.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.q_norm": false, "model.layers.29.self_attn.k_norm": false, "model.layers.29.self_attn.rope": false, "model.layers.29.input_layernorm": false, "model.layers.29.post_attention_layernorm": false, "model.layers.29.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.30.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.q_norm": false, "model.layers.30.self_attn.k_norm": false, "model.layers.30.self_attn.rope": false, "model.layers.30.input_layernorm": false, "model.layers.30.post_attention_layernorm": false, "model.layers.30.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.q_norm": false, "model.layers.31.self_attn.k_norm": false, "model.layers.31.self_attn.rope": false, "model.layers.31.input_layernorm": false, "model.layers.31.post_attention_layernorm": false, "model.layers.31.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.32.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.q_norm": false, "model.layers.32.self_attn.k_norm": false, "model.layers.32.self_attn.rope": false, "model.layers.32.input_layernorm": false, "model.layers.32.post_attention_layernorm": false, "model.layers.32.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.33.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.q_norm": false, "model.layers.33.self_attn.k_norm": false, "model.layers.33.self_attn.rope": false, "model.layers.33.input_layernorm": false, "model.layers.33.post_attention_layernorm": false, "model.layers.33.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.q_norm": false, "model.layers.34.self_attn.k_norm": false, "model.layers.34.self_attn.rope": false, "model.layers.34.input_layernorm": false, "model.layers.34.post_attention_layernorm": false, "model.layers.34.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.35.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.q_norm": false, "model.layers.35.self_attn.k_norm": false, "model.layers.35.self_attn.rope": false, "model.layers.35.input_layernorm": false, "model.layers.35.post_attention_layernorm": false, "model.layers.35.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.36.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.q_norm": false, "model.layers.36.self_attn.k_norm": false, "model.layers.36.self_attn.rope": false, "model.layers.36.input_layernorm": false, "model.layers.36.post_attention_layernorm": false, "model.layers.36.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.q_norm": false, "model.layers.37.self_attn.k_norm": false, "model.layers.37.self_attn.rope": false, "model.layers.37.input_layernorm": false, "model.layers.37.post_attention_layernorm": false, "model.layers.37.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.38.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.q_norm": false, "model.layers.38.self_attn.k_norm": false, "model.layers.38.self_attn.rope": false, "model.layers.38.input_layernorm": false, "model.layers.38.post_attention_layernorm": false, "model.layers.38.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.39.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.q_norm": false, "model.layers.39.self_attn.k_norm": false, "model.layers.39.self_attn.rope": false, "model.layers.39.input_layernorm": false, "model.layers.39.post_attention_layernorm": false, "model.layers.39.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.q_norm": false, "model.layers.40.self_attn.k_norm": false, "model.layers.40.self_attn.rope": false, "model.layers.40.input_layernorm": false, "model.layers.40.post_attention_layernorm": false, "model.layers.40.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.41.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.q_norm": false, "model.layers.41.self_attn.k_norm": false, "model.layers.41.self_attn.rope": false, "model.layers.41.input_layernorm": false, "model.layers.41.post_attention_layernorm": false, "model.layers.41.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.42.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.42.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.q_norm": false, "model.layers.42.self_attn.k_norm": false, "model.layers.42.self_attn.rope": false, "model.layers.42.input_layernorm": false, "model.layers.42.post_attention_layernorm": false, "model.layers.42.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.43.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.43.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.q_norm": false, "model.layers.43.self_attn.k_norm": false, "model.layers.43.self_attn.rope": false, "model.layers.43.input_layernorm": false, "model.layers.43.post_attention_layernorm": false, "model.layers.43.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.44.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.44.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.q_norm": false, "model.layers.44.self_attn.k_norm": false, "model.layers.44.self_attn.rope": false, "model.layers.44.input_layernorm": false, "model.layers.44.post_attention_layernorm": false, "model.layers.44.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.45.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.45.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.q_norm": false, "model.layers.45.self_attn.k_norm": false, "model.layers.45.self_attn.rope": false, "model.layers.45.input_layernorm": false, "model.layers.45.post_attention_layernorm": false, "model.layers.45.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.46.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.46.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.q_norm": false, "model.layers.46.self_attn.k_norm": false, "model.layers.46.self_attn.rope": false, "model.layers.46.input_layernorm": false, "model.layers.46.post_attention_layernorm": false, "model.layers.46.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.47.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.47.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.q_norm": false, "model.layers.47.self_attn.k_norm": false, "model.layers.47.self_attn.rope": false, "model.layers.47.input_layernorm": false, "model.layers.47.post_attention_layernorm": false, "model.layers.47.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.norm": false, "lm_head": { "group_size": 64, "bits": 4 } }, "quantization_config": { "group_size": 64, "bits": 4, "model.embed_tokens": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.0.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.self_attn.q_norm": false, "model.layers.0.self_attn.k_norm": false, "model.layers.0.self_attn.rope": false, "model.layers.0.input_layernorm": false, "model.layers.0.post_attention_layernorm": false, "model.layers.0.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.0.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.1.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.1.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.self_attn.q_norm": false, "model.layers.1.self_attn.k_norm": false, "model.layers.1.self_attn.rope": false, "model.layers.1.input_layernorm": false, "model.layers.1.post_attention_layernorm": false, "model.layers.1.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.1.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.2.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.2.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.self_attn.q_norm": false, "model.layers.2.self_attn.k_norm": false, "model.layers.2.self_attn.rope": false, "model.layers.2.input_layernorm": false, "model.layers.2.post_attention_layernorm": false, "model.layers.2.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.2.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.3.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.3.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.self_attn.q_norm": false, "model.layers.3.self_attn.k_norm": false, "model.layers.3.self_attn.rope": false, "model.layers.3.input_layernorm": false, "model.layers.3.post_attention_layernorm": false, "model.layers.3.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.3.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.4.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.4.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.self_attn.q_norm": false, "model.layers.4.self_attn.k_norm": false, "model.layers.4.self_attn.rope": false, "model.layers.4.input_layernorm": false, "model.layers.4.post_attention_layernorm": false, "model.layers.4.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.4.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.5.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.5.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.self_attn.q_norm": false, "model.layers.5.self_attn.k_norm": false, "model.layers.5.self_attn.rope": false, "model.layers.5.input_layernorm": false, "model.layers.5.post_attention_layernorm": false, "model.layers.5.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.5.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.6.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.self_attn.q_norm": false, "model.layers.6.self_attn.k_norm": false, "model.layers.6.self_attn.rope": false, "model.layers.6.input_layernorm": false, "model.layers.6.post_attention_layernorm": false, "model.layers.6.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.6.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.self_attn.q_norm": false, "model.layers.7.self_attn.k_norm": false, "model.layers.7.self_attn.rope": false, "model.layers.7.input_layernorm": false, "model.layers.7.post_attention_layernorm": false, "model.layers.7.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.7.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.8.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.self_attn.q_norm": false, "model.layers.8.self_attn.k_norm": false, "model.layers.8.self_attn.rope": false, "model.layers.8.input_layernorm": false, "model.layers.8.post_attention_layernorm": false, "model.layers.8.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.8.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.9.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.self_attn.q_norm": false, "model.layers.9.self_attn.k_norm": false, "model.layers.9.self_attn.rope": false, "model.layers.9.input_layernorm": false, "model.layers.9.post_attention_layernorm": false, "model.layers.9.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.9.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.self_attn.q_norm": false, "model.layers.10.self_attn.k_norm": false, "model.layers.10.self_attn.rope": false, "model.layers.10.input_layernorm": false, "model.layers.10.post_attention_layernorm": false, "model.layers.10.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.10.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.11.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.self_attn.q_norm": false, "model.layers.11.self_attn.k_norm": false, "model.layers.11.self_attn.rope": false, "model.layers.11.input_layernorm": false, "model.layers.11.post_attention_layernorm": false, "model.layers.11.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.11.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.12.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.self_attn.q_norm": false, "model.layers.12.self_attn.k_norm": false, "model.layers.12.self_attn.rope": false, "model.layers.12.input_layernorm": false, "model.layers.12.post_attention_layernorm": false, "model.layers.12.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.12.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.self_attn.q_norm": false, "model.layers.13.self_attn.k_norm": false, "model.layers.13.self_attn.rope": false, "model.layers.13.input_layernorm": false, "model.layers.13.post_attention_layernorm": false, "model.layers.13.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.13.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.14.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.self_attn.q_norm": false, "model.layers.14.self_attn.k_norm": false, "model.layers.14.self_attn.rope": false, "model.layers.14.input_layernorm": false, "model.layers.14.post_attention_layernorm": false, "model.layers.14.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.14.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.15.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.self_attn.q_norm": false, "model.layers.15.self_attn.k_norm": false, "model.layers.15.self_attn.rope": false, "model.layers.15.input_layernorm": false, "model.layers.15.post_attention_layernorm": false, "model.layers.15.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.15.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.self_attn.q_norm": false, "model.layers.16.self_attn.k_norm": false, "model.layers.16.self_attn.rope": false, "model.layers.16.input_layernorm": false, "model.layers.16.post_attention_layernorm": false, "model.layers.16.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.16.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.17.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.self_attn.q_norm": false, "model.layers.17.self_attn.k_norm": false, "model.layers.17.self_attn.rope": false, "model.layers.17.input_layernorm": false, "model.layers.17.post_attention_layernorm": false, "model.layers.17.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.17.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.18.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.self_attn.q_norm": false, "model.layers.18.self_attn.k_norm": false, "model.layers.18.self_attn.rope": false, "model.layers.18.input_layernorm": false, "model.layers.18.post_attention_layernorm": false, "model.layers.18.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.18.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.self_attn.q_norm": false, "model.layers.19.self_attn.k_norm": false, "model.layers.19.self_attn.rope": false, "model.layers.19.input_layernorm": false, "model.layers.19.post_attention_layernorm": false, "model.layers.19.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.19.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.20.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.self_attn.q_norm": false, "model.layers.20.self_attn.k_norm": false, "model.layers.20.self_attn.rope": false, "model.layers.20.input_layernorm": false, "model.layers.20.post_attention_layernorm": false, "model.layers.20.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.20.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.21.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.self_attn.q_norm": false, "model.layers.21.self_attn.k_norm": false, "model.layers.21.self_attn.rope": false, "model.layers.21.input_layernorm": false, "model.layers.21.post_attention_layernorm": false, "model.layers.21.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.21.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.self_attn.q_norm": false, "model.layers.22.self_attn.k_norm": false, "model.layers.22.self_attn.rope": false, "model.layers.22.input_layernorm": false, "model.layers.22.post_attention_layernorm": false, "model.layers.22.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.22.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.23.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.self_attn.q_norm": false, "model.layers.23.self_attn.k_norm": false, "model.layers.23.self_attn.rope": false, "model.layers.23.input_layernorm": false, "model.layers.23.post_attention_layernorm": false, "model.layers.23.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.23.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.24.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.self_attn.q_norm": false, "model.layers.24.self_attn.k_norm": false, "model.layers.24.self_attn.rope": false, "model.layers.24.input_layernorm": false, "model.layers.24.post_attention_layernorm": false, "model.layers.24.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.24.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.self_attn.q_norm": false, "model.layers.25.self_attn.k_norm": false, "model.layers.25.self_attn.rope": false, "model.layers.25.input_layernorm": false, "model.layers.25.post_attention_layernorm": false, "model.layers.25.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.25.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.26.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.self_attn.q_norm": false, "model.layers.26.self_attn.k_norm": false, "model.layers.26.self_attn.rope": false, "model.layers.26.input_layernorm": false, "model.layers.26.post_attention_layernorm": false, "model.layers.26.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.26.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.27.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.self_attn.q_norm": false, "model.layers.27.self_attn.k_norm": false, "model.layers.27.self_attn.rope": false, "model.layers.27.input_layernorm": false, "model.layers.27.post_attention_layernorm": false, "model.layers.27.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.27.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.self_attn.q_norm": false, "model.layers.28.self_attn.k_norm": false, "model.layers.28.self_attn.rope": false, "model.layers.28.input_layernorm": false, "model.layers.28.post_attention_layernorm": false, "model.layers.28.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.28.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.29.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.self_attn.q_norm": false, "model.layers.29.self_attn.k_norm": false, "model.layers.29.self_attn.rope": false, "model.layers.29.input_layernorm": false, "model.layers.29.post_attention_layernorm": false, "model.layers.29.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.29.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.30.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.self_attn.q_norm": false, "model.layers.30.self_attn.k_norm": false, "model.layers.30.self_attn.rope": false, "model.layers.30.input_layernorm": false, "model.layers.30.post_attention_layernorm": false, "model.layers.30.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.30.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.self_attn.q_norm": false, "model.layers.31.self_attn.k_norm": false, "model.layers.31.self_attn.rope": false, "model.layers.31.input_layernorm": false, "model.layers.31.post_attention_layernorm": false, "model.layers.31.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.31.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.32.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.self_attn.q_norm": false, "model.layers.32.self_attn.k_norm": false, "model.layers.32.self_attn.rope": false, "model.layers.32.input_layernorm": false, "model.layers.32.post_attention_layernorm": false, "model.layers.32.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.32.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.33.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.self_attn.q_norm": false, "model.layers.33.self_attn.k_norm": false, "model.layers.33.self_attn.rope": false, "model.layers.33.input_layernorm": false, "model.layers.33.post_attention_layernorm": false, "model.layers.33.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.33.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.self_attn.q_norm": false, "model.layers.34.self_attn.k_norm": false, "model.layers.34.self_attn.rope": false, "model.layers.34.input_layernorm": false, "model.layers.34.post_attention_layernorm": false, "model.layers.34.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.34.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.35.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.self_attn.q_norm": false, "model.layers.35.self_attn.k_norm": false, "model.layers.35.self_attn.rope": false, "model.layers.35.input_layernorm": false, "model.layers.35.post_attention_layernorm": false, "model.layers.35.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.35.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.36.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.self_attn.q_norm": false, "model.layers.36.self_attn.k_norm": false, "model.layers.36.self_attn.rope": false, "model.layers.36.input_layernorm": false, "model.layers.36.post_attention_layernorm": false, "model.layers.36.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.36.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.self_attn.q_norm": false, "model.layers.37.self_attn.k_norm": false, "model.layers.37.self_attn.rope": false, "model.layers.37.input_layernorm": false, "model.layers.37.post_attention_layernorm": false, "model.layers.37.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.37.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.38.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.self_attn.q_norm": false, "model.layers.38.self_attn.k_norm": false, "model.layers.38.self_attn.rope": false, "model.layers.38.input_layernorm": false, "model.layers.38.post_attention_layernorm": false, "model.layers.38.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.38.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.39.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.self_attn.q_norm": false, "model.layers.39.self_attn.k_norm": false, "model.layers.39.self_attn.rope": false, "model.layers.39.input_layernorm": false, "model.layers.39.post_attention_layernorm": false, "model.layers.39.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.39.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.v_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.self_attn.q_norm": false, "model.layers.40.self_attn.k_norm": false, "model.layers.40.self_attn.rope": false, "model.layers.40.input_layernorm": false, "model.layers.40.post_attention_layernorm": false, "model.layers.40.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.40.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.41.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.self_attn.q_norm": false, "model.layers.41.self_attn.k_norm": false, "model.layers.41.self_attn.rope": false, "model.layers.41.input_layernorm": false, "model.layers.41.post_attention_layernorm": false, "model.layers.41.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.41.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.42.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.42.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.self_attn.q_norm": false, "model.layers.42.self_attn.k_norm": false, "model.layers.42.self_attn.rope": false, "model.layers.42.input_layernorm": false, "model.layers.42.post_attention_layernorm": false, "model.layers.42.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.42.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.43.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.43.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.self_attn.q_norm": false, "model.layers.43.self_attn.k_norm": false, "model.layers.43.self_attn.rope": false, "model.layers.43.input_layernorm": false, "model.layers.43.post_attention_layernorm": false, "model.layers.43.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.43.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.44.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.44.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.self_attn.q_norm": false, "model.layers.44.self_attn.k_norm": false, "model.layers.44.self_attn.rope": false, "model.layers.44.input_layernorm": false, "model.layers.44.post_attention_layernorm": false, "model.layers.44.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.44.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.45.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.45.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.self_attn.q_norm": false, "model.layers.45.self_attn.k_norm": false, "model.layers.45.self_attn.rope": false, "model.layers.45.input_layernorm": false, "model.layers.45.post_attention_layernorm": false, "model.layers.45.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.45.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.46.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.46.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.self_attn.q_norm": false, "model.layers.46.self_attn.k_norm": false, "model.layers.46.self_attn.rope": false, "model.layers.46.input_layernorm": false, "model.layers.46.post_attention_layernorm": false, "model.layers.46.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.46.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.layers.47.self_attn.q_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.k_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.v_proj": { "group_size": 64, "bits": 4 }, "model.layers.47.self_attn.o_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.self_attn.q_norm": false, "model.layers.47.self_attn.k_norm": false, "model.layers.47.self_attn.rope": false, "model.layers.47.input_layernorm": false, "model.layers.47.post_attention_layernorm": false, "model.layers.47.mlp.gate": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.gate_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.up_proj": { "group_size": 64, "bits": 3 }, "model.layers.47.mlp.switch_mlp.down_proj": { "group_size": 64, "bits": 4 }, "model.norm": false, "lm_head": { "group_size": 64, "bits": 4 } }, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "router_aux_loss_coef": 0.001, "sliding_window": null, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "4.51.0", "use_cache": true, "use_sliding_window": false, "vocab_size": 151936 }