benjaminfh
commented
10 months ago I have had this - I can't remember the exact cause but I'm fairly sure it's a transformers version issue. I'd first check that you're trying to load your model into an env with the same transformers version as the env you did training in. Also check version requirements in general - circa 4.25.1 works best in my experience.
I brain dumped my experience in python module version management here a few days ago. It might be useful: https://github.com/clovaai/donut/issues/132#issuecomment-1837611702
Hi Donut team,
I trained "donut-base" model with my custom dataset for testing but I cannot do inference that it returned error.
################ Could not find image processor class in the image processor config or the model config. Loading based on pattern matching with the model's feature extractor configuration.
You are using a model of type donut to instantiate a model of type vision-encoder-decoder. This is not supported for all configurations of models and can yield errors.
Traceback (most recent call last): File "C:\Users\Admin\projects\training\donut\inference.py", line 19, in
processor, model = load_model(model_id)
File "C:\Users\Admin\projects\training\donut\inference.py", line 16, in load_model
model = VisionEncoderDecoderModel.from_pretrained(model_id)
File "C:\Users\Admin\anaconda3\envs\aienv\lib\site-packages\transformers\models\vision_encoder_decoder\modeling_vision_encoder_decoder.py", line 363, in from_pretrained
return super().from_pretrained(pretrained_model_name_or_path, *model_args, kwargs)
File "C:\Users\Admin\anaconda3\envs\aienv\lib\site-packages\transformers\modeling_utils.py", line 2535, in from_pretrained
config, model_kwargs = cls.config_class.from_pretrained(
File "C:\Users\Admin\anaconda3\envs\aienv\lib\site-packages\transformers\configuration_utils.py", line 598, in from_pretrained
return cls.from_dict(config_dict, kwargs)
File "C:\Users\Admin\anaconda3\envs\aienv\lib\site-packages\transformers\configuration_utils.py", line 747, in from_dict
config = cls(**config_dict)
File "C:\Users\Admin\anaconda3\envs\aienv\lib\site-packages\transformers\models\vision_encoder_decoder\configuration_vision_encoder_decoder.py", line 85, in init
raise ValueError(
ValueError: A configuraton of type donut cannot be instantiated because not both
encoderanddecodersub-configurations are passed, but only {'_name_or_path': './models/donut-base', 'align_long_axis': False, 'architectures': ['DonutModel'], 'decoder_layer': 4, 'encoder_layer': [2, 2, 14, 2], 'input_size': [1280, 960], 'max_length': 768, 'max_position_embeddings': 768, 'model_type': 'donut', 'torch_dtype': 'float32', 'transformers_version': '4.33.3', 'window_size': 10, '_commit_hash': None} ################Could you please give an advice for this case?
Thanks & Regards, Felix