Exception: data did not match any variant of untagged enum ModelWrapper at line 646524 column 3

#2
by Maverick17 - opened

I can't run your code. The tokenizer name should be "deepseek-ai/deepseek-vl2", I guess.

---> 12 vl_chat_processor: DeepseekVLV2Processor = DeepseekVLV2Processor.from_pretrained(model_path)
     13 tokenizer = vl_chat_processor.tokenizer
     14 

/usr/local/lib/python3.10/dist-packages/transformers/processing_utils.py in from_pretrained(cls, pretrained_model_name_or_path, cache_dir, force_download, local_files_only, token, revision, **kwargs)
    890         # Determine if kwargs is a flat dictionary or contains nested dictionaries
    891         if any(key in default_kwargs for key in kwargs):
--> 892             # kwargs is dictionary-based, and some keys match modality names
    893             for modality, subdict in kwargs.items():
    894                 if modality in default_kwargs:

/usr/local/lib/python3.10/dist-packages/transformers/processing_utils.py in _get_arguments_from_pretrained(cls, pretrained_model_name_or_path, **kwargs)
    936 
    937         </Tip>
--> 938 
    939         Args:
    940             pretrained_model_name_or_path (`str` or `os.PathLike`):

/usr/local/lib/python3.10/dist-packages/transformers/tokenization_utils_base.py in from_pretrained(cls, pretrained_model_name_or_path, cache_dir, force_download, local_files_only, token, revision, trust_remote_code, *init_inputs, **kwargs)
   2269         for key in cls.SPECIAL_TOKENS_ATTRIBUTES & init_kwargs.keys():
   2270             if added_tokens_map != {} and init_kwargs[key] is not None:
-> 2271                 if key != "additional_special_tokens":
   2272                     init_kwargs[key] = added_tokens_map.get(str(init_kwargs[key]), init_kwargs[key])
   2273 

/usr/local/lib/python3.10/dist-packages/transformers/tokenization_utils_base.py in _from_pretrained(cls, resolved_vocab_files, pretrained_model_name_or_path, init_configuration, token, cache_dir, local_files_only, _commit_hash, _is_local, trust_remote_code, *init_inputs, **kwargs)
   2503             save_directory=save_directory,
   2504             file_names=file_names,
-> 2505             legacy_format=legacy_format,
   2506             filename_prefix=filename_prefix,
   2507         )

/usr/local/lib/python3.10/dist-packages/transformers/models/llama/tokenization_llama_fast.py in __init__(self, vocab_file, tokenizer_file, clean_up_tokenization_spaces, unk_token, bos_token, eos_token, add_bos_token, add_eos_token, use_default_system_prompt, legacy, add_prefix_space, **kwargs)
    155             kwargs["from_slow"] = True
    156 
--> 157         super().__init__(
    158             vocab_file=vocab_file,
    159             tokenizer_file=tokenizer_file,

/usr/local/lib/python3.10/dist-packages/transformers/tokenization_utils_fast.py in __init__(self, *args, **kwargs)
    113             fast_tokenizer = copy.deepcopy(tokenizer_object)
    114         elif fast_tokenizer_file is not None and not from_slow:
--> 115             # We have a serialization from tokenizers which let us directly build the backend
    116             fast_tokenizer = TokenizerFast.from_file(fast_tokenizer_file)
    117         elif slow_tokenizer:

Exception: data did not match any variant of untagged enum ModelWrapper at line 646524 column 3

Sign up or log in to comment