tokenizer_config.json

{
  "tokenizer_class": "PreTrainedTokenizerFast",
  "model_max_length": 512,
  "padding_side": "right",
  "truncation_side": "right",
  "clean_up_tokenization_spaces": false,
  "bos_token": "<|start|>",
  "eos_token": "<|sep|>",
  "sep_token": "<|sep|>",
  "cls_token": "<|cls|>",
  "unk_token": "<|unk|>",
  "pad_token": "<|pad|>",
  "mask_token": "<|mask|>",
  "add_prefix_space": false
}