Refactor FNet and CVT output tracing

[beelapranay]

What does this PR do?

Refactors FNet and CVT output tracing to use the standardized decorators where appropriate.

1. FNet now uses @capture_outputs with _can_record_outputs to collect hidden states.
2. CVT keeps manual hidden-state collection to preserve vision-model semantics.
3. Wrapper heads use @can_return_tuple to simplify return_dict handling.

Fixes #43979

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. Link
• 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?

@molbap @ArthurZucker

[github-actions]

[For maintainers] Suggested jobs to run (before merge)

run-slow: cvt, fnet