tokenizer_config.json

{
  "add_prefix_space": false,
  "additional_special_tokens": [
    "<SYSTEM_TASK:>",
    "<USER_TASK:>",
    "<ASSISTANT_TASK:>",
    "<END_TASK>"
  ],
  "bos_token": "<|endoftext|>",
  "clean_up_tokenization_spaces": true,
  "eos_token": "<|endoftext|>",
  "model_max_length": 3000,
  "pad_token": "<PAD>",
  "tokenizer_class": "CodeGenTokenizer",
  "unk_token": "<|endoftext|>",
  "vocab_size": 50300
}