About Mutil-Gpus for training?

[Nan.Y]

Hi, authors:

I found an issue when training the code with multi-GPUs, which is that only a single GPU is used despite inputting multiple GPUs. Can you solve this problem in your spare time?

Thanks!

[Xinyang Li]

Hi! The code is expected to support multi-gpu training now with DataParallel.
Can you share which command you used?

[Nan.Y]

I train the model with your code and your official command, i.e., python core/train.py --config configs/celeba-hq.yaml --gpus 0,1.
However, only one single GPU is used when training. The reason for this issue may be that Class "Gen" has no function of forward(self, ..). Can you help me?

[Nan.Y]

A link for this issue is in here . Can you solve this problem in your spare time? @imlixinyang
Thanks!

[Xinyang Li]

Try command "python core/train.py --config configs/celeba-hq.yaml --gpus 0 1".

[Nan.Y]

I will try it now. Thanks for your reply. I am doing a new work based on your novel work. I will cite your work.
Thanks for your hard work.

[Nan.Y]

It works now. Thanks again. @imlixinyang

[Xinyang Li]

Glad to hear that! Gook luck for your research.