Classification recipe stdout insights are poluted by spammy warning about interpolation argument

[Horgix]

When running the Classification recipe as per the hello world, the stdout helpful insights are diluted in the middle of the same warning again and again:

[...]/python3.9/site-packages/torchvision/transforms/functional.py:364: UserWarning: Argument interpolation should be of type InterpolationMode instead of int. Please, use InterpolationMode enum.
  warnings.warn(

I believe this is a bug, and makes the output hard to read (80% garbage)