{
  "architectures": [
    "RITAModel"
  ],
  "bos_token_id": [
    50256
  ],
  "d_feedforward": 4096,
  "d_model": 1024,
  "dropout": 0.0,
  "eos_token_id": 50256,
  "max_seq_len": 1024,
  "model_type": "codegen",
  "num_heads": 16,
  "num_layers": 24,
  "torch_dtype": "float32",
  "transformers_version": "4.16.2",
  "vocab_size": 128
}