model_config.json

{
    "model_class": "VisualChatModel",
    "tokenizer_type": "Meta-Llama-3-8B-Instruct",
    "num_layers": 32,
    "hidden_size": 4096,
    "num_attention_heads": 32,
    "vocab_size": 128256,
    "layernorm_order": "pre",
    "model_parallel_size": 1,
    "max_sequence_length": 8192,
    "use_bias": false,
    "inner_hidden_size": 14336,
    "num_multi_query_heads": 8,
    "image_length": 2304,
    "image_size": 1344,
    "eva_args": {
        "model_class": "EVA2CLIPModel",
        "num_layers": 63,
        "hidden_size": 1792,
        "num_attention_heads": 16,
        "vocab_size": 1,
        "layernorm_order": "post",
        "model_parallel_size": 1,
        "max_sequence_length": 257,
        "inner_hidden_size": 15360,
        "use_final_layernorm": false,
        "layernorm_epsilon": 1e-06,
        "row_parallel_linear_final_bias": false,
        "image_size": [
            1344,
            1344
        ],
        "pre_len": 1,
        "post_len": 0,
        "in_channels": 3,
        "patch_size": 14
    },
    "bos_token_id": 128000,
    "eos_token_id": 128001,
    "pad_token_id": null
}