{
    "architectures": [
        "Ernie4_5_MoeForCausalLM"
    ],
    "bos_token_id": 1,
    "dtype": "bfloat16",
    "eos_token_id": 2,
    "hidden_act": "silu",
    "hidden_size": 2560,
    "initializer_range": 0.02,
    "intermediate_size": 12288,
    "max_position_embeddings": 131072,
    "model_type": "ernie4_5_moe",
    "moe_capacity": [
        64,
        64,
        64
    ],
    "moe_gate": "top2_fused",
    "moe_intermediate_size": 1536,
    "moe_k": 6,
    "moe_layer_end_index": 47,
    "moe_layer_interval": 1,
    "moe_layer_start_index": 1,
    "moe_norm_min": 1e-12,
    "moe_num_experts": 64,
    "moe_num_shared_experts": 2,
    "moe_use_aux_free": true,
    "num_attention_heads": 20,
    "num_hidden_layers": 47,
    "num_key_value_heads": 4,
    "output_router_logits": false,
    "pad_token_id": 0,
    "quantization": {
        "group_size": 32,
        "bits": 6,
        "mode": "affine",
        "model.embed_tokens": {
            "group_size": 32,
            "bits": 4
        },
        "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.down_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.0.mlp.up_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.1.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.1.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.1.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.1.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.2.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.2.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.2.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.3.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.3.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.3.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.4.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.4.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.4.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "lm_head": {
            "group_size": 32,
            "bits": 6
        }
    },
    "quantization_config": {
        "group_size": 32,
        "bits": 6,
        "mode": "affine",
        "model.embed_tokens": {
            "group_size": 32,
            "bits": 4
        },
        "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.down_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.0.mlp.up_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.1.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.1.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.1.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.1.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.1.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.2.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.2.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.2.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.2.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.3.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.3.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.3.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.3.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.self_attn.v_proj": {
            "group_size": 32,
            "bits": 8
        },
        "model.layers.4.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.4.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.4.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.4.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.5.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.6.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.7.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.7.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.8.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.9.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.10.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.10.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.11.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.12.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.13.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.13.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.14.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.15.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.16.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.16.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.17.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.18.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.19.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.19.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.20.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.21.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.22.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.22.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.23.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.24.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.25.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.25.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.26.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.27.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.28.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.28.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.29.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.30.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.31.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.31.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.32.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.33.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.34.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.34.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.35.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.36.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.37.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.37.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.38.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.v_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.39.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.40.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.40.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.41.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.41.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.42.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.42.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.43.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.43.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.44.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.44.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.45.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.45.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.q_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.k_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.self_attn.v_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.self_attn.o_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.gate": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.switch_mlp.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.mlp.shared_experts.gate_proj": {
            "group_size": 32,
            "bits": 4
        },
        "model.layers.46.mlp.shared_experts.down_proj": {
            "group_size": 32,
            "bits": 6
        },
        "model.layers.46.mlp.shared_experts.up_proj": {
            "group_size": 32,
            "bits": 4
        },
        "lm_head": {
            "group_size": 32,
            "bits": 6
        }
    },
    "rms_norm_eps": 1e-05,
    "rope_scaling": null,
    "rope_theta": 500000,
    "router_aux_loss_coef": 0.001,
    "tie_word_embeddings": false,
    "transformers_version": "4.56.1",
    "use_bias": false,
    "use_cache": true,
    "vocab_size": 103424
}