galactica-orca-wizardlm-1.3b/tokenizer_config.json

{
  "clean_up_tokenization_spaces": true,
  "model_max_length": 2048,
  "tokenizer_class": "PreTrainedTokenizerFast"
}