Drop inplace operation for loss computation with gradient accumulation

[qgallouedec]

What does this PR do?

Fix

RuntimeError: a leaf Variable that requires grad is being used in an in-place operation.

reported here: https://huggingface.co/answerdotai/ModernBERT-base/discussions/9#6769e0390519344317755b3b

Initially introduced in #35207

The solution does solve the issue, tested locally.

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?

@muellerzr @SunMarc @ArthurZucker who reviewed #35207

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[qgallouedec]

Not related, but relate to the same lines: #35207 (comment) @yzhangcs Ignore: point raised wrong

[SunMarc]

Nice thanks for fixing it so quickly ! I'll merge this as others are out for a while and this is safe to merge !