config.json

{
  "architectures": ["LLaMAForCausalLM"],
  "model_type": "llama",
  "vocab_size": 32000,
  "hidden_size": 4096,
  "num_hidden_layers": 32,
  "num_attention_heads": 32,
  "intermediate_size": 11008,
  "max_position_embeddings": 2048,
  "layer_norm_eps": 1e-5,
  "initializer_range": 0.02,
  "hidden_act": "swiglu",
  "rope_scaling": {
    "type": "linear",
    "factor": 1.0
  },
  "use_cache": true
}