Why type torch.FloatTensor but got torch.cuda.FloatTensor?

sefibk / KernelGAN

Other

340 stars 77 forks source link

Why type torch.FloatTensor but got torch.cuda.FloatTensor? #1

Closed zhuwanling closed 4 years ago

zhuwanling commented 5 years ago

Thank you for sharing. I get a error: RuntimeError: Function AddBackward0 returned an invalid gradient at index 1 - expected type torch.FloatTensor but got torch.cuda.FloatTensor In # Calculate gradientstotal_loss_g.backward() We have made any changes.

sefibk commented 4 years ago

Sorry but I am not sure. Did you follow the .yml file precisely?