config.json

{
  "_name_or_path": "meta-llama/Meta-Llama-3.1-8B-Instruct",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 128000,
  "eos_token_id": [
    128001,
    128008,
    128009
  ],
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 14336,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.45.1",
  "use_cache": true,
  "vocab_size": 128256,
  "quantization_config": {
    "quant_method": "vptq",
    "shared_layer_config":{
      "down_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 14336,
        "in_features": 14336,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 4096,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "gate_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 14336,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "up_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 14336,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "k_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 1024,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "o_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 4096,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "q_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 4096,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      },
      "v_proj": {
        "bias": false,
        "enable_norm": true,
        "enable_perm": false,
        "group_num": 1,
        "group_size": 4096,
        "in_features": 4096,
        "indices_as_float": false,
        "is_indice_packed": true,
        "num_centroids": [
          -1,
          65536
        ],
        "num_res_centroids": [
          -1,
          4096
        ],
        "out_features": 1024,
        "outlier_size": 0,
        "vector_lens": [
          -1,
          12
        ]
      }
    }
  }
}