{
  "name_or_path": "xlm-roberta-base",
  "special_tokens_map_file": null,
  "tokenizer_class": "PreTrainedTokenizerFast",
  "use_fast": true
}