tokenizer_config.json

{
  "model_max_length": 2048,
  "padding_side": "left",
  "special_tokens_map_file": "special_tokens_map.json",
  "tokenizer_class": "PreTrainedTokenizerFast"
}