{ "architectures": [ "Glm4MoeForCausalLM" ], "attention_bias": true, "attention_dropout": 0.0, "eos_token_id": [ 151329, 151336, 151338 ], "first_k_dense_replace": 1, "head_dim": 128, "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 10944, "max_position_embeddings": 131072, "model_type": "glm4_moe", "moe_intermediate_size": 1408, "n_group": 1, "n_routed_experts": 128, "n_shared_experts": 1, "norm_topk_prob": true, "num_attention_heads": 96, "num_experts_per_tok": 8, "num_hidden_layers": 46, "num_key_value_heads": 8, "num_nextn_predict_layers": 1, "pad_token_id": 151329, "partial_rotary_factor": 0.5, "quantization": { "group_size": 32, "bits": 8, "mode": "affine", "model.embed_tokens": { "group_size": 32, "bits": 6 }, "model.layers.0.self_attn.q_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.k_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.o_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.gate_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.up_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.5.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.7.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.7.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.8.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.10.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.10.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.11.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.13.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.13.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.14.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.16.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.16.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.17.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.19.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.19.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.20.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.22.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.22.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.23.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.25.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.25.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.26.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.28.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.28.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.29.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.31.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.31.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.32.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.34.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.34.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.35.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.37.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.37.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.38.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.40.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.40.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "lm_head": { "group_size": 32, "bits": 8 } }, "quantization_config": { "group_size": 32, "bits": 8, "mode": "affine", "model.embed_tokens": { "group_size": 32, "bits": 6 }, "model.layers.0.self_attn.q_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.k_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.self_attn.o_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.gate_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.up_proj": { "group_size": 32, "bits": 8 }, "model.layers.0.mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.1.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.1.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.2.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.2.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.3.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.3.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.4.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.4.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.5.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.5.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.6.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.7.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.7.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.7.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.8.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.8.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.9.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.10.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.10.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.10.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.11.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.11.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.12.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.13.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.13.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.13.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.14.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.14.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.15.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.16.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.16.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.16.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.17.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.17.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.18.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.19.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.19.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.19.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.20.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.20.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.21.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.22.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.22.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.22.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.23.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.23.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.24.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.25.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.25.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.25.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.26.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.26.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.27.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.28.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.28.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.28.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.29.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.29.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.30.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.31.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.31.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.31.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.32.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.32.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.33.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.34.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.34.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.34.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.35.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.35.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.36.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.37.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.37.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.37.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.38.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.38.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.v_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.39.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.40.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.40.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.40.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.41.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.41.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.42.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.42.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.43.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.43.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.44.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.44.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.self_attn.q_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.self_attn.k_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.self_attn.v_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.self_attn.o_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.switch_mlp.down_proj": { "group_size": 32, "bits": 8 }, "model.layers.45.mlp.shared_experts.gate_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.shared_experts.up_proj": { "group_size": 32, "bits": 6 }, "model.layers.45.mlp.shared_experts.down_proj": { "group_size": 32, "bits": 8 }, "lm_head": { "group_size": 32, "bits": 8 } }, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 1000000, "routed_scaling_factor": 1.0, "tie_word_embeddings": false, "topk_group": 1, "torch_dtype": "bfloat16", "transformers_version": "4.54.0", "use_cache": true, "use_qk_norm": false, "vocab_size": 151552 }