Add support for async checkpointing

[rohitgr7]

What does this PR do?

Fixes #11561

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

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 @awaelchli @ananthsub @ninginthecloud @rohitgr7 @otaj @akihironitta

[rohitgr7]

after discussion with @carmocca

the design

class AsyncCheckpointIO(CheckpointIO):
    def __init__(self, checkpoint_io=None):
        self._checkpoint_io = checkpoint_io

    @setter
    def checkpoint_io(...):
        ...

    def save_checkpoint(self, checkpoint, filepath)
        async_launch(self._checkpoint_io.save_checkpoint, checkpoint, filepath)

and in the Trainer, the user can just pass

Trainer(plugins=[AsyncCheckpointIO()])

we will attach the appropriate checkpoint if not already attached. This will enable async checkpointing for existing plugins by providing async behavior without any code modification.

[codecov]

Codecov Report

Merging #13658 (c129206) into master (6a9d626) will decrease coverage by 10%.
The diff coverage is 82%.

@@            Coverage Diff            @@
##           master   #13658     +/-   ##
=========================================
- Coverage      86%      76%    -10%     
=========================================
  Files         330      332      +2     
  Lines       25973    26048     +75     
=========================================
- Hits        22266    19808   -2458     
- Misses       3707     6240   +2533     

[awaelchli]

Great work @rohitgr7!!

[rohitgr7]

Thanks @awaelchli @carmocca @otaj for great reviews and discussion.