tokenizer_config.json

{
  "vocab_size": 1000,
  "model_max_length": 64,
  "tokenizer_class": "SimpleVocabTokenizer"
}