Pluralistic-Inpainting
Pluralistic-Inpainting copied to clipboard
Resize/crop
Hi, When you train the model, how do you resize the ImageNet images to 256x256 ? Your code suggests that you resize first to 266x266 and then crop. Is it true ? Do you do that for other datasets too that are already 256x256 ?
Hi @samuro95, thanks for your interest. Yes, we directly resize images to a larger size and then crop them to 256x256 on all datasets. This is a general data augmentation method on deep learning. During the testing, we will directly resize images to 256x256.