[Python] Should the shuffle=True argument be used during training in...

I am confused regarding whether the shuffle=True argument should be used in the pytorch torch.utils.data.DataLoader() during training when the training data are time series data. I believe that shuffle=True should be used during training in most cases, but does this apply to time series data (where the order matters) as well?

Continue reading...