tokenizer_config.json

{
  "tokenizer_class": "BertTokenizer",
  "do_lower_case": false,
  "model_max_length": 256,
  "tokenizer_type": "neuralmind/bert-base-portuguese-cased"
}