tokenizer_config.json

{
  "added_tokens_decoder": {},
  "additional_special_tokens": [
    "<|startoftext|>",
    "<|extra_0|>",
    "<|extra_4|>",
    "<|extra_5|>",
    "<|eos|>"
  ],
  "architectures": [
    "GPT2LMHeadModel"
  ],
  "auto_map": {
    "AutoTokenizer": [
      "tokenization_hy.HYTokenizer",
      null
    ]
  },
  "bos_token": "<|startoftext|>",
  "clean_up_tokenization_spaces": false,
  "eos_token": "<|eos|>",
  "extra_special_tokens": {},
  "model_max_length": 262144,
  "model_type": "gpt2",
  "pad_token": "<|pad|>",
  "padding_side": "right",
  "split_special_tokens": false,
  "tokenizer_class": "HYTokenizer"
}