Train error due to detach

shleecs / DeRaindrop_unofficial

Unofficial training part implementation: Attentive Generative Adversarial Network for Raindrop Removal from A Single Image (CVPR 2018)

24 stars 8 forks source link

Train error due to detach #2

Open dister520 opened 4 years ago

dister520 commented 4 years ago

I am trying to train and am seeing the following error. File "/home/bhaktivideo/Downloads/DeRaindrop_unofficial-master/train_module.py", line 139, in train_start loss_G, loss_D, loss_PL, loss_ML, loss_att, loss_MAP, MSE_loss= self.forwardprocess(I,GT_) File "/home/bhaktivideo/Downloads/DeRaindrop_unofficial-master/train_module.py", line 87, in forward_process lossPL = self.criterionPL(O, GT_.detach())

I did add autograd to the system. Any advice thank you?

Sun-Happy-YKX commented 1 year ago

i meets the same problem as well, could you please share your solutions?