jxm

/

cde-small-v2

like 74

Env: Transformers version: 4.48.1(current newest)

2 problems:

1. raise TypeError: TypeError: no_grad.init() takes 1 positional argument but 2 were given

---------------------------------------------------------------------------TypeError Traceback (most recent call last)
Cell In[2], line 1----> 1 model = transformers.AutoModel.from_pretrained("/data/yy/llm/cde_model/cde-small-v2", trust_remote_code=True)
File ~/miniconda3/envs/yi_dev/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py:526, in _BaseAutoModelClass.from_pretrained(cls, pretrained_model_name_or_path, *model_args, **kwargs) 523 if kwargs.get("quantization_config", None) is not None: 524 _ = kwargs.pop("quantization_config")--> 526 config, kwargs = AutoConfig.from_pretrained( 527 pretrained_model_name_or_path, 528 return_unused_kwargs=True, 529 trust_remote_code=trust_remote_code, 530 code_revision=code_revision, 531 _commit_hash=commit_hash, 532 **hub_kwargs, 533 **kwargs, 534 ) 536 # if torch_dtype=auto was passed here, ensure to pass it on 537 if kwargs_orig.get("torch_dtype", None) == "auto":

File ~/miniconda3/envs/yi_dev/lib/python3.10/site-packages/transformers/models/auto/configuration_auto.py:1063, in AutoConfig.from_pretrained(cls, pretrained_model_name_or_path, **kwargs) 1061 if has_remote_code and trust_remote_code: 1062 class_ref = config_dict["auto_map"]["AutoConfig"]-> 1063 config_class = get_class_from_dynamic_module( 1064 class_ref, pretrained_model_name_or_path, code_revision=code_revision, **kwargs 1065 ) 1066 if os.path.isdir(pretrained_model_name_or_path): 1067 config_class.register_for_auto_class()

File ~/miniconda3/envs/yi_dev/lib/python3.10/site-packages/transformers/dynamic_module_utils.py:553, in get_class_from_dynamic_module(class_reference, pretrained_model_name_or_path, cache_dir, force_download, resume_download, proxies, token, revision, local_files_only, repo_type, code_revision, **kwargs) 540 # And lastly we get the class inside our newly created module 541 final_module = get_cached_module_file( 542 repo_id, 543 module_file + ".py", (...) 551 repo_type=repo_type, 552 )--> 553 return get_class_in_module(class_name, final_module, force_reload=force_download)
File ~/miniconda3/envs/yi_dev/lib/python3.10/site-packages/transformers/dynamic_module_utils.py:250, in get_class_in_module(class_name, module_path, force_reload) 248 # reload in both cases, unless the module is already imported and the hash hits 249 if getattr(module, "transformers_module_hash", "") != module_hash:--> 250 module_spec.loader.exec_module(module) 251 module.transformers_module_hash = module_hash 252 return getattr(module, class_name)

File :883, in exec_module(self, module)
File :241, in _call_with_frames_removed(f, *args, **kwds)
File ~/.cache/huggingface/modules/transformers_modules/cde-small-v2/model.py:269 264 else: 265 return t[min_row:max_row] 268 @torch .no_grad--> 269 def maxsim( 270 X: torch.Tensor, y: torch.Tensor, 271 maximize: bool, chunk_size: int = 8_000, 272 debug_mem_usage: bool = False) -> torch.Tensor: 273 device = X.device 274 n_samples = X.shape[0]TypeError: no_grad.init() takes 1 positional argument but 2 were given

The key issue occurs at line 268 of model.py:
268 @torch .no_grad
269 def maxsim(...

If I change the @torch .no_grad into @torch .no_grad() followed by copilot's suggestion, then the second problem showed up:

2. forcely download answerdotai/ModernBERT-base model, and if my internet connection works bad, downloading will desperately fail.
   I'm wondering if Either jxm/cde-small-v2 and answerdotai/ModernBERT-base are essential, where could I place the answerdotai/ModernBERT-base model wich is downloaded manually?

I have tried to have the ModernBERT-base folder including config.json etc being the child folder of the set cde-small-v2 folder, but that doesn't work.