tokenizer/tokenizer_config.json

{
  "vocab_size": 32000,
  "max_seq_len": 128,
  "pad_id": 0,
  "cls_id": 2,
  "sep_id": 3,
  "pad_token": "[PAD]",
  "cls_token": "[CLS]",
  "sep_token": "[SEP]"
}