enhance: avoid unnecessary fallback in _bincount on CUDA with deterministic

[hyukkyukang]

What does this PR do?

Update _bincount to use native torch.bincount on CUDA when deterministic algorithms are enabled (PyTorch ≥ 2.1), as this use case is now officially supported. Retain the fallback implementation for MPS, XLA, and older PyTorch versions. Also improved the docstring to reflect updated backend support and trade-offs.

Fixes #3086

Before submitting

• Was this discussed/agreed via a Github issue? (no need for typos and docs improvements)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure to update the docs?
• Did you write any new necessary tests?

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

---

📚 Documentation preview 📚: https://torchmetrics--3087.org.readthedocs.build/en/3087/

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 70%. Comparing base (74bdb26) to head (6b1e16a).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master   #3087    +/-   ##
=======================================
- Coverage      70%     70%    -0%     
=======================================
  Files         359     345    -14     
  Lines       19892   19702   -190     
=======================================
- Hits        13931   13742   -189     
+ Misses       5961    5960     -1     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Borda]

@hyukkyukang seems that the GPU tests are failing for a reason related to your change...
for PT 2.7 and CUDA 12.3
RuntimeError: called a synchronizing CUDA operation