Merge different gpu backends with accelerator='gpu'

[justusschock]

What does this PR do?

Setting accelerator='gpu' does no longer select the GPUAccelerator (see #13636 ), but will dynamically choose an available GPU backend (currently cuda or mps)

Fixes #13102

Does your PR introduce any breaking changes? If yes, please list them.

Yes, together with #13636 this changes accelerator='gpu' to dynamically select gpu backends

Before submitting

• Was this discussed/approved via a GitHub issue? (not for typos and docs)
• 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? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

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:

• 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

Did you have fun?

Make sure you had fun coding 🙃

cc @Borda @justusschock @kaushikb11 @awaelchli @akihironitta @rohitgr7

[codecov]

Codecov Report

Merging #13642 (a0c76b9) into master (8d14554) will increase coverage by 27%.
The diff coverage is 100%.

❗ Current head a0c76b9 differs from pull request most recent head 92de866. Consider uploading reports for the commit 92de866 to get more accurate results

@@            Coverage Diff            @@
##           master   #13642     +/-   ##
=========================================
+ Coverage      49%      76%    +27%     
=========================================
  Files         328      327      -1     
  Lines       25552    25679    +127     
=========================================
+ Hits        12470    19538   +7068     
+ Misses      13082     6141   -6941     

[carmocca]

@justusschock CI failures look real

[awaelchli]

Fixed it. After the changes we did for ddp fork #13405, we should mock the device parser functions instead of torch.cuda.device_count directly.