[Python] How can I load a pretrained transformers model that was manually downloaded?

I am unable to download huggingface models through the Python functions due to SSL certificate errors. Perhaps it's due to my company firewall.

I am able to download the contents of a huggingface model repo through a browser to a folder. I'm trying to load this model from disk using TFPreTrainedModel.from_pretrained() and AutoTokenizer.from_pretrained(). From the docs, it seems this is a valid option.

I'm receiving an error message that isn't useful AttributeError: 'NoneType' object has no attribute 'from_pretrained'. Appreciate any help!

Example repo:

https://huggingface.co/cardiffnlp/twitter-roberta-base-sentiment-latest/tree/main

Code

from transformers import pipeline, TFPreTrainedModel, AutoTokenizer
import os

dir = "./models/twitter-roberta-base-sentiment-latest/"
print(os.listdir(dir)) # confirm the folder contents

model = TFPreTrainedModel.from_pretrained(dir)
tokenizer = AutoTokenizer.from_pretrained(dir)

analyze = pipeline(task="sentiment-analyis", model=model, tokenizer=tokenizer)
print(analyze("this is good"))
print(analyze("this is bad"))


Output

2025-02-21 16:40:05.896448: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2025-02-21 16:40:06.653841: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
WARNING:tensorflow:From C:\Users\xxxxx\.pyenv\pyenv-win\versions\3.12.8\Lib\site-packages\tf_keras\src\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.

['config.json', 'gitattributes', 'merges.txt', 'pytorch_model.bin', 'README.md', 'special_tokens_map.json', 'tf_model.h5', 'vocab.json']
Traceback (most recent call last):
File "C:\Users\xxxxx\OneDrive - DuPont\Python Projects\huggingface\sentiment.py", line 8, in <module>
model = TFPreTrainedModel.from_pretrained(dir)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\xxxxx\.pyenv\pyenv-win\versions\3.12.8\Lib\site-packages\transformers\modeling_tf_utils.py", line 2726, in from_pretrained
config, model_kwargs = cls.config_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'NoneType' object has no attribute 'from_pretrained'


Docs

https://huggingface.co/docs/transfo..._classes/model#transformers.TFPreTrainedModel

pretrained_model_name_or_path (str, optional) — Can be either: A string, the model id of a pretrained model hosted inside a model repo on huggingface.co. A path to a directory containing model weights saved using save_pretrained(), e.g., ./my_model_directory/. A path or url to a PyTorch state_dict save file (e.g, ./pt_model/pytorch_model.bin). In this case, from_pt should be set to True and a configuration object should be provided as config argument. This loading path is slower than converting the PyTorch model in a TensorFlow model using the provided conversion scripts and loading the TensorFlow model afterwards. None if you are both providing the configuration and state dictionary (resp. with keyword arguments config and state_dict).

Continue reading...