config.json

{
  "architectures": [
    "LlamaForCausalLM"
  ],
  "model_type": "llama",
  "torch_dtype": "float16",
  "transformers_version": "4.44.0",
  "task_specific_params": {
    "text-generation": {}
  }
}