Fix Trainer MLU Availability Import For Accelerate < 0.29

[alex-jw-brooks]

What does this PR do?

Fixes #32461

I checked Accelerate, and the utility for checking if MLU is available was added in the 0.29.0 release, so the accelerate import in the trainer breaks versions 0.21.0 >= x <0.29.0 if accelerate is available at import time (since the minimum version of accelerate transformers is happy with is 0.21.0 in the deps table).

It appears that the is_mlu_available check in the Accelerate import utils (here) is also only used one time and is redundant with the is_torch_mlu_available check in transformers import utils here, which is used in other parts of the trainer.

This PR removes the accelerate util that is breaking the import, and just uses the transformers one in its place.

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?

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

@muellerzr and @SunMarc

[alex-jw-brooks]

This is fixed by the changes in #32292