tokenizer_config.json

{
  "clean_up_tokenization_spaces": false,
  "model_max_length": 65536,
  "tokenizer_class": "PreTrainedTokenizerFast",
  "unk_token": "<unk>"
}