fix: AdvancedProfiler no longer crashes when stop() is called after teardown

[pooyanazad]

What does this PR do?

This PR fixes a bug where AdvancedProfiler crashes with a ValueError if a profiling context manager is active during trainer teardown.

When teardown() is called, it clears the profiled_actions dictionary. Exiting the context manager subsequently calls stop(), which was crashing because the action no longer existed.
I fixed this by:

1. Making stop() return silently if the action doesn't exist.
2. Disabling all active cProfile instances in teardown() before clearing the dictionary to prevent global profiler leaks.

Fixes #9136

Before submitting

• Was this discussed/agreed via a GitHub issue? Yes, issue AdvancedProfiler: ValueError: Attempting to stop recording an action (run_test_evaluation) which was never started. #9136.
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (N/A)
• Did you write any new necessary tests? Yes, added regression tests in test_profiler.py.
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request? (No breaking changes)
• Did you update the CHANGELOG? (N/A - minor bugfix)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified

[pooyanazad]

Hi, we need a reviewer here