{
  "architectures": [
    "GPTLMHeadModel"
  ],
  "model_type": "gpt2",
  "vocab_size": 8000,
  "n_positions": 512,
  "n_embd": 768,
  "n_layer": 12,
  "n_head": 12,
  "n_inner": 3072,
  "activation_function": "gelu",
  "layer_norm_epsilon": 1e-05,
  "tokenizer_class": "Tokenizer",
  "torch_dtype": "float32"
}