Value Error when instantiating MultiClass Classifier

Here is the code I am using:
model = ClassificationModel('roberta', 'roberta-base', num_labels=37, args={'learning_rate':1e-5, 'num_train_epochs':10, 'reprocess_input_data':True})

Here is the error it generates:

ValueError                                Traceback (most recent call last)
<ipython-input-16-affadfcfb25e> in <module>
      1 #Need to figure out how many labels in os data
----> 2 model = ClassificationModel('roberta', 'roberta-base', num_labels=37, args={'learning_rate':1e-5, 'num_train_epochs':10, 'reprocess_input_data':True})

~\Anaconda3\lib\site-packages\simpletransformers\classification\ in __init__(self, model_type, model_name, tokenizer_type, tokenizer_name, num_labels, weight, args, use_cuda, cuda_device, onnx_execution_provider, **kwargs)
    335         if num_labels:
--> 336             self.config = config_class.from_pretrained(
    337                 model_name, num_labels=num_labels, **self.args.config
    338             )

~\Anaconda3\lib\site-packages\transformers\ in from_pretrained(cls, pretrained_model_name_or_path, **kwargs)
    426         """
--> 427         config_dict, kwargs = cls.get_config_dict(pretrained_model_name_or_path, **kwargs)
    428         if "model_type" in config_dict and hasattr(cls, "model_type") and config_dict["model_type"] != cls.model_type:
    429             logger.warn(

~\Anaconda3\lib\site-packages\transformers\ in get_config_dict(cls, pretrained_model_name_or_path, **kwargs)
    482         try:
    483             # Load from URL or cache if already cached
--> 484             resolved_config_file = cached_path(
    485                 config_file,
    486                 cache_dir=cache_dir,

~\Anaconda3\lib\site-packages\transformers\ in cached_path(url_or_filename, cache_dir, force_download, proxies, resume_download, user_agent, extract_compressed_file, force_extract, use_auth_token, local_files_only)
   1269     if is_remote_url(url_or_filename):
   1270         # URL, so get it from the cache (downloading if necessary)
-> 1271         output_path = get_from_cache(
   1272             url_or_filename,
   1273             cache_dir=cache_dir,

~\Anaconda3\lib\site-packages\transformers\ in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, use_auth_token, local_files_only)
   1492                     )
   1493                 else:
-> 1494                     raise ValueError(
   1495                         "Connection error, and we cannot find the requested files in the cached path."
   1496                         " Please try again or make sure your Internet connection is on."

ValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.

What can I do to resolve this issue?

hey @parkz your issue seems to be about the simpletransformers library (not transformers) so i suggest you post you question on their repo to see if someone can help: GitHub - ThilinaRajapakse/simpletransformers: Transformers for Classification, NER, QA, Language Modelling, Language Generation, T5, Multi-Modal, and Conversational AI

1 Like

Is there a guide to setting up a multiclass with transformers?

yes, there’s an official guide for text classification here: 🤗 Transformers Notebooks — transformers 4.5.0.dev0 documentation

in that notebook you can see how the number of labels is specified in the model config and you could try adapting this code to your use case :slight_smile: