training on multiple GPUs

[enes3774]

firstly thanks for such an amazing and clear implementation of the paper. I want to train the model on 8 GPUs with 24GB. Should ı add only torch.nn.DataParallel or should ı change other codes as well? Thanks in advance

[jhgan00]

"I apologize for the delayed response. Currently, I am unable to test the code in a multi-GPU environment. Therefore, it is difficult for me to provide a definite answer, but it seems that DataParallel should be able to work. Thank you."