tokenizer_config.json

{
  "model_max_length": 1000000000000000019884624838656,
  "name_or_path": "./vi_tokenizer.json",
  "tokenizer_class": "PreTrainedTokenizerFast"
}