tokenizer_config.json

{
  "vocab_size": 16384,
  "endoftext_id": 0,
  "pad_id": 1,
  "endoftext_token": "<|endoftext|>",
  "pad_token": "<|pad|>"
}