consideragain

Thanks for your reply. The versions are as follows: pytorch --------------- 1.8.1 torchvision----------- 0.9.1 opencv-python------4.5.2.54 cudatoolkit-----------10.2.89

What is your Python and Ubuntu versions?

I'm not sure if I understand you correctly. I changed this line to "padding = [0, 0] + list(np.maximum(self.padding - np.array(img.size), 0))" but this is invalid. There is the same...

My Ubuntu was reinstalled. So I can't try the idea. Now I give up reappearing this code. Thank for your help.