Update setup logic in training type plugins [1 / n]

[awaelchli]

What does this PR do?

Part of #9987

Adds methods for plugins to override how models and optimizers are setup.

The follow up PR(s) will add implementations for setup to DP, TPUSpawn, Deepspeed, etc.

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

No.

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 update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read 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?

I made sure I had fun coding 🙃

Make sure you had fun coding 🙃

[SeanNaren]

I do resonate with @four4fish, if we'd like to get this in before our release, we should try to comment/todo the introduced function which should be refactored once the precision plugin is moved into the TTP (I think out of scope for 1.5)

[awaelchli]

@SeanNaren @four4fish Thanks for the suggestions! Added docs and TODO

[tchaton]

LGTM 1

[four4fish]

@awaelchli @tchaton @SeanNaren Thank you for the explanations!