tokenizer_config.json

{
  "add_prefix_space": false,
  "eos_token": "<end_message>",
  "model_max_length": 8192,
  "special_tokens_map_file": null,
  "tokenizer_class": "PreTrainedTokenizerFast"
}