wonwoongcho / gdwct Goto Github PK

When I set the DATA_PARALLEL True, I met a runtime error: arguments are located on different GPUs. Anyone knows how to solve it? Thx

when is the plan to release pretrained models?

Hello!
Great paper! I am trying to train on a custom dataset, after one iteration the training just stops, any idea why?

Firstly, Thank you for your great job.
After I download celebA dataset using download.sh, how to preprocess celebA dataset? (i.e. split trainA and trainB and so on)
Could you upload the preprocessing code or tell me web site which teaches the preprocessing method?
Thanks.

I tried to train on my custom datasets from scratch.
However, the generated image is so bad as follow.

Do your datasets have to be the same size? Or do they just have to be divisible by batch size? I was trying this repo out but ran into some issues where randomly the losses would become NaN.

My trainA has 33344 images and my trainB has 22933 images. All hyperparams were default except dataset paths/names.

Thank you for your great job!

I cannot use your code for research because your repository does not include a license file (i.e. no license).
Please add license information.
Thanks.

Thanks for the great work! My question is that how the hyperparameters are set as it is a complex network to train?