tokenizer_config.json

{
  "tokenizer_type": "tiktoken",
  "vocab_size": 50261,
  "model": "gpt2"
}