I want to use the UNet output image, but I get 2 different images, can tell me the reason?
first, use this code ‘torchvision.utils.save_image(img, os.path.join(result_path, str(i + 1) + ‘image.bmp’))’
second,use this code ’ im = torch.squeeze(im)
im = transforms.ToPILImage()(im)
im_resized = im.resize((100, 100))
!
im_resized.save(os.path.join(result_path, str(i+1) + ‘SR.bmp’)) ’
What is the difference in both images?
Do they look completely different?
Did you try to load both after saving and compare their values?
How large is the difference?