Add PyTorch version check for FA backend on AMD GPUs

[mht-sharma]

What does this PR do?

Updates condition from PR #30850

Garbage values may occur during model generation with models like LLama, Mistral, and Mixtral, particularly when utilizing multi-gpu setup and device_map=auto alongside SDPA and FA.

The above issue is fixed in the newer PyTorch versions.

Fixes # (issue)

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?

[Rocketknight1]

cc @SunMarc @MekkCyber @muellerzr

[MekkCyber]

Good catch ! LGTM, thanks for the fix !

[SunMarc]

SG