Getting issue while loading model about bert-sentiment HOT 6 CLOSED

did you use data parallel to train the model?

from bert-sentiment.

@abhishekkrthakur Yes used data parallel while training

from bert-sentiment.

does your bert_base_path has bert base uncased model files?

from bert-sentiment.

@abhishekkrthakur Fixed the issue was with data parallel. I local i was not making use of "MODEL = nn.DataParallel(MODEL)". Now working with this. Can you help me understand the use of DataParallel and if it does require after training also?
Thanks for quick replying. Good to close the issue

from bert-sentiment.

DataParallel is used only when you have multiple GPUs during training.
If you used it in training, you have to use in inference but there are other ways too.

Closing this issue for now. :)

from bert-sentiment.

@abhishekkrthakur : Can you give any leads on how to load DataParallel GPU model on CPU?
As per pytorch docs tried following but still raises above RuntimeError

device = torch.device('cpu')
model = TheModelClass(*args, **kwargs)
model.load_state_dict(torch.load(PATH, map_location=device))

from bert-sentiment.