tokenizer_config (9).json

{
  "clean_up_tokenization_spaces": true,
  "model_max_length": 1024,
  "tokenizer_class": "EsmTokenizer"
}