{
  "model_type": "llama",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "tokenizer_class": "TokenizersBackend",
  "model_class": "LlamaForCausalLM",
  "vocab_size": 128000,
  "params": 1260505088,
  "torch_dtype": "torch.float16"
}