tokenizer_config.json

{
  "bos_token": "<|startoftext|>",
  "clean_up_tokenization_spaces": true,
  "do_clean_text": false,
  "eos_token": "<|endoftext|>",
  "model_max_length": 1280,
  "pad_token": "<|endoftext|>",
  "sep_token": "<|segmenter|>",
  "special_tokens_map_file": null,
  "tokenizer_class": "GPTSanJapaneseTokenizer",
  "unk_token": "<|endoftext|>"
}