CodeGeeX2模型转换错误 about chatglm.cpp HOT 2 OPEN

执行： python3 chatglm_cpp/convert.py -i modules/codegeex2-6b -t q4_0 -o codegeex-ggml.bin 报错： Traceback (most recent call last): File "chatglm_cpp/convert.py", line 543, in <module> main() File "chatglm_cpp/convert.py", line 537, in main convert(f, args.model_name_or_path, dtype=args.type) File "chatglm_cpp/convert.py", line 469, in convert tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True) File "/Users/eric/miniconda3/envs/chatglm.cpp/lib/python3.8/site-packages/transformers/models/auto/tokenization_auto.py", line 774, in from_pretrained return tokenizer_class.from_pretrained(pretrained_model_name_or_path, *inputs, **kwargs) File "/Users/eric/miniconda3/envs/chatglm.cpp/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 2028, in from_pretrained return cls._from_pretrained( File "/Users/eric/miniconda3/envs/chatglm.cpp/lib/python3.8/site-packages/transformers/tokenization_utils_base.py", line 2260, in _from_pretrained tokenizer = cls(*init_inputs, **init_kwargs) File "/Users/eric/.cache/huggingface/modules/transformers_modules/codegeex2-6b/tokenization_chatglm.py", line 69, in __init__ super().__init__(padding_side=padding_side, clean_up_tokenization_spaces=False, **kwargs) File "/Users/eric/miniconda3/envs/chatglm.cpp/lib/python3.8/site-packages/transformers/tokenization_utils.py", line 367, in __init__ self._add_tokens( File "/Users/eric/miniconda3/envs/chatglm.cpp/lib/python3.8/site-packages/transformers/tokenization_utils.py", line 467, in _add_tokens current_vocab = self.get_vocab().copy() File "/Users/eric/.cache/huggingface/modules/transformers_modules/codegeex2-6b/tokenization_chatglm.py", line 99, in get_vocab vocab = {self._convert_id_to_token(i): i for i in range(self.vocab_size)} File "/Users/eric/.cache/huggingface/modules/transformers_modules/codegeex2-6b/tokenization_chatglm.py", line 95, in vocab_size return self.tokenizer.n_words AttributeError: 'ChatGLMTokenizer' object has no attribute 'tokenizer'

#190 (comment)

from chatglm.cpp.

chatchat-space/Langchain-Chatchat#1835 (comment)

这两个方式都能解决问题，要么降级，要么使用 chatgml3-6b 下的 文件 把 codegeex2-6b 下的覆盖掉

from chatglm.cpp.