{
    "architectures": [
        "LlamaForCausalLM"
    ],
    "model_type": "llama",
    "hidden_size": 4096,
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "intermediate_size": 16384,
    "max_position_embeddings": 512,
    "vocab_size": 50257,
    "type_vocab_size": 1,
    "initializer_range": 0.02,
    "layer_norm_eps": 1e-12,
    "pad_token_id": 50256,
    "bos_token_id": 50256,
    "eos_token_id": 50256,
    "gradient_checkpointing": false
}