CenterCrop after Resize with the same size value

[tolgacangoz]

Hi diffusers community,
I wonder what is the reason for using transforms.CenterCrop(size) after transforms.Resize(size) with the same values. Isn't CenterCrop(size) unnecessary? Example.

[asomoza]

size is an int in that script:

parser.add_argument(
        "--resolution",
        type=int,
        default=512,
        help=(
            "The resolution for input images, all the images in the train/validation dataset will be resized to this"
            " resolution"
        ),
    )

transforms.Resize(args.resolution)

If size is an int, smaller edge of the image will be matched to this number. i.e, if height > width, then image will be rescaled to (size * height / width, size).

transforms.CenterCrop(args.resolution)

If size is an int instead of sequence like (h, w), a square crop (size, size) is made.

So in short, Is necessary if the image is not a square image, resize maintains the aspect ratio of the image.

[tolgacangoz]

Hmm, you are right, thank you 👍.