[NEW MODEL] MViTV2

[kamila-chay]

What does this PR do?

This PR introduces MViT, which is the basis for MeMViT, which is a video transformer that got requested in the issue below. I'm currently doing research in fine-grained video understanding and MeMViT happens to be able to process long video sequences. Unfortunately it's only available from the oficial FAIR repo which can be hard to integrate. That's why I decided to implement it here, but to make things cleaner and more modular I started with MViTv2 which is also a great vision transformer (mainly for image-related tasks)

Fixes 20545

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

@amyeroberts @qubvel