{
  "model_type": "MiniArt",
  "architectures": ["MiniArtForCausalLM"],
  "tokenizer_class": "GPT2Tokenizer",
  "bos_token": "",
  "eos_token": "<|im_end|>",
  "context_length": 32768,
  "model_max_length": 32768
}