text_tokenizer/added_tokens.json

{
  "<extra_id_0>": 250100,
  "<extra_id_1>": 250101,
  "<extra_id_2>": 250102
}