{
  "note": "This model uses GPT-2 tokenizer. Please use: tokenizer = GPT2Tokenizer.from_pretrained('gpt2')",
  "vocab_size": 50257,
  "encoding": "gpt2"
}