tokenizer_config.json

{
  "add_prefix_space": false,
  "backend": "tokenizers",
  "bos_token": "<|begin|>",
  "clean_up_tokenization_spaces": false,
  "eos_token": "<|end|>",
  "errors": "replace",
  "extra_special_tokens": [
    "<|doc|>",
    "<|system|>",
    "<|user|>",
    "<|assistant|>"
  ],
  "is_local": true,
  "local_files_only": false,
  "model_max_length": 1000000000000,
  "pad_token": "<|pad|>",
  "tokenizer_class": "GPT2Tokenizer",
  "unk_token": "<|endoftext|>",
  "vocab_size": 49152
}