I am getting the following error when I try to load my saved Pytorch model.
ModuleNotFoundError: No module named 'transformers.models'
The way I am loading the model is : model = torch.load(r'Models-BERT\model-duplicates dropped(sampled)_v2', map_location=torch.device('cpu'))
I saved the model using: torch.save(model, 'Models-BERT/model-duplicates dropped(sampled)_v2')
The version of transformers used: 3.4.0
also tested on transformers version 4.12.2
So far I have tried upgrading the transformers library, I have tried the solution provided in the similar question on StackOverflow here:StackOverflow similar question but nothing worked. I also ran my code on Azure notebook and on my local system, both are giving same error.
When I run the same snippet on Azure, I am getting
ModuleNotFoundError: No module named 'transformers.modeling_bert'
What could be the issue here? And how can I run this code?
