Fix attention backend leaking across models

[raghulchandramouli]

What does this PR do?

In a nutshell, It Prevents ModelMixin.set_attention_backend() from changing the process-global active attention backend. The method now only configures attention processors belonging to the target model.

Self-review:

Ran self-review and these are the validation I got from it:

• Blocking issues: None
• Non-blocking issues: None
• Dead code: None
• Documentation changes required: None

Test I Ran:

• Regression tests for Flux and QwenImage: all passed
• make style: passed
• make fix-copies: passed
• git diff --check: passed

Test I wrote:

• Added a regression test in tests/attention.py to make sure that ModelMixin.set_attention_backend() does not modify the registry.

discussion

This appears as an issue raised by sayak Paul - #14249

Fixes #14249

Before submitting

• Did you use an AI agent (Claude Code, Codex, Cursor, etc.) to help with this PR? If so:
  • Did you read the Coding with AI agents guide?
  • Did you run the self-review skill on the diff?
  • Did you share the final self-review notes in the PR description or a comment?
• Did you read the contributor guideline?
• Did you read our philosophy doc? (important for complex PRs)
• 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?
• Are you the author (or part of the team) of the model/pipeline (only applicable for model/pipeline related PRs)?

It would be great if @qgallouedec or @sayakpaul can you take a quick review of this PR??