{
  "architectures": [
    "LlamaForCausalLM"
  ],
  "model_type": "llama",
  "hidden_size": 768,
  "num_hidden_layers": 6,
  "num_attention_heads": 4,
  "vocab_size": 16384,
  "intermediate_size": 2048,
  "rms_norm_eps": 1e-06,
  "max_position_embeddings": 4096,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "pad_token_id": 2,
  "torch_dtype": "float32",
  "tie_word_embeddings": true
}