tokenizer_config.json

{
  "unk_token": "<|end_of_text|>",
  "sep_token": "<|end_of_text|>",
  "pad_token": "<|end_of_text|>",
  "cls_token": "<|begin_of_text|>",
  "mask_token": null,
  "bos_token": "<|begin_of_text|>",
  "eos_token": "<|end_of_text|>",
  "additional_special_tokens": ["<|eom_id|>", "<|eot_id|>", "<|python_tag|>"],
  "model_max_length": 2048,
  "tokenizer_class": "PreTrainedTokenizerFast",
  "clean_up_tokenization_spaces": true
}