tileb1

Hi @gwenzek, I also tried the branch and it works with my singularity image. It would be great to have it merged!

Ah yes, didn't think of loading from the checkpoint... Thanks!

Ok great, thank you @shijianjian !

@shijianjian, I don't get the error anymore now. Though the parameters cannot be optimized. The following code is taken as is from the paper. ``` import kornia.augmentation as K; import...

@shijianjian then, I get: ``` tensor([0., 0.]) tensor([0., 0.]) ``` So now at least I have gradient, but they are 0 for some reason. If you check the loss I...

Hey @armanhar, thanks for the reply! Can you let me know once you have a timeline for this feature? Thanks

@armanhar great, thanks!

Hi @nartmada, this issue was created almost 2 years ago and I don't have this bug anymore with the more recent PyTorch/rocm version.