enable arbitrary batch size for all prototype kernels

[pmeier]

Fixes #6670. This basically only moves the squashing that we did for the video kernels into the image kernels.

If we start to merge transforms v1 and v2, we should support this better. I haven't checked, but I guesstimate that 90% - 100% of all times we need to squash / unsquash is because we use these three ops internally:

1. vision/torchvision/transforms/functional_tensor.py

   Line 469 in 6e203b4

   img = interpolate(img, size=size, mode=interpolation, align_corners=align_corners, antialias=antialias)

2. vision/torchvision/transforms/functional_tensor.py

   Line 562 in 6e203b4

   img = grid_sample(img, grid, mode=mode, padding_mode="zeros", align_corners=False)

3. vision/torchvision/transforms/functional_tensor.py

   Line 763 in 6e203b4

   img = conv2d(img, kernel, groups=img.shape[-3])

I propose, we implement _nd_{interpolate, grid_sample, conv_2d} as internal utilities, so we don't have to have the arbitrary batch compatibility code everywhere. This is somewhat already implemented for 2. with

vision/torchvision/transforms/functional_tensor.py

Line 547 in 6e203b4

def _apply_grid_transform(

[datumbox]

LGTM, thanks!

FYI, you are missing an import. Feel fee to merge on green CI.

[pmeier]

FYI, you are missing an import.

Well, no. It was somewhat worse than that: I accidentally removed the perspective dispatcher, but I put it back now 😛