tokenizer_config.json

{
  "tokenizer_class": "GPT2Tokenizer",
  "vocab_size": 50257,
  "model_max_length": 512,
  "bos_token": "<|endoftext|>",
  "eos_token": "<|endoftext|>",
  "unk_token": "<|endoftext|>",
  "pad_token": "<|endoftext|>",
  "add_prefix_space": false,
  "do_lower_case": false
}