Bug in clip resizing

[aderylo]

There seems to be a bug in the Resizing code:

vjepa2/src/datasets/utils/video/functional.py

Line 75 in c2963a4

size = size[1], size[0]

the swap is unnecessary and it should be size = size[0], size[1] given that docstring of Resize class states:

    """Resizes a list of (H x W x C) numpy.ndarray to the final size
    The larger the original image is, the more times it takes to
    interpolate
    Args:
    interpolation (str): Can be one of 'nearest', 'bilinear'
    defaults to nearest
    size (tuple): (width, height)
    """

and PILs Image.resize expects also size to be (width, height) not (height, width) as it is now.