AdvSemiSeg icon indicating copy to clipboard operation
AdvSemiSeg copied to clipboard

Why use trainloader_iter and trainloader_gt_iter at the same time?

Open daifeng2016 opened this issue 4 years ago • 4 comments

Hi, I am wondering why you use trainloader_iter and trainloader_gt_iter at the same time during your training process, since they are both refer to the labeled dataset. During the D training, only labels_gt is used from trainloader_gt_iter, what is the difference if using labels from trainloader_gt_iter?

daifeng2016 avatar Dec 21 '19 05:12 daifeng2016

It's simply for generating two sampling sequences for the entire dataset. one for the "real" branch and one for the fake "branch". You can also sample twice from the same data loader. We only do this for better readability since data loader in this project does not cost much efficiency.

hfslyc avatar Jan 16 '20 23:01 hfslyc

@hfslyc Hi, can I replace the labels_gt with the labels? I mean if I can just sample once from the same trainloader_iter. Does is it matter?

songkq avatar Nov 06 '20 12:11 songkq

@ipScore I think it doesn't matter.

hfslyc avatar Nov 07 '20 07:11 hfslyc

@hfslyc Thanks.

songkq avatar Nov 07 '20 08:11 songkq