Vectorize RepetitionPenaltyLogitsProcessor to improve performance

[bdalal]

What does this PR do?

This PR replaces the nested loops in the RepetitionPenaltyLogitsProcessor with a vectorized implementation to provide speedups on long sequences of roughly 3 orders of magnitude on GPUs and 2 orders of magnitude on CPUs.

Fixes # 8596

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 the 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?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors which may be interested in your PR.

@patrickvonplaten

[bdalal]

ValueError: Connection error, and we cannot find the requested files in the cached path. Please try again or make sure your Internet connection is on.

Seems like a flake. Tests passes locally.

[patrickvonplaten]

This is really great work @bdalal ! Tested in on equivalence even if input_ids contain same ids => and it's much faster, great!

Merging.