tokenizer_config.json

{
  "tokenizer_class": "CharLevelTokenizer",
  "unk_token": "[UNK]"
}