tokenizer_config.json

{
  "additional_special_tokens": null,
  "backend": "tokenizers",
  "model_input_names": ["input_ids", "attention_mask"],
  "bos_token": "<|start|>",
  "cls_token": "<|cls|>",
  "eos_token": "<|end|>",
  "is_local": false,
  "mask_token": "<|mask|>",
  "model_max_length": 1024,
  "pad_token": "<|pad|>",
  "padding_side": "right",
  "sep_token": "<|sep|>",
  "special_tokens_map": {
    "bos_token": "<|start|>",
    "cls_token": "<|cls|>",
    "eos_token": "<|end|>",
    "mask_token": "<|mask|>",
    "pad_token": "<|pad|>",
    "sep_token": "<|sep|>",
    "unk_token": "<|unk|>"
  },
  "tokenizer_class": "PreTrainedTokenizerFast",
  "truncation": "longest_first",
  "unk_token": "<|unk|>"
}