tokenizer_config.json

{
  "vocab_size": 50000,
  "min_frequency": 2,
  "model_type": "byte_level_bpe",
  "special_tokens": {
    "eos_token": "<|endoftext|>",
    "pad_token": "<|pad|>",
    "unk_token": "<|unk|>",
    "mask_token": "<|mask|>"
  }
}