Allow users to specify custom fit_loop in Trainer

[lematt1991]

What does this PR do?

This PR adds fit_loop to the Trainer constructor, allowing users to supply custom fit loops. This would solve #2288, allowing users to train full batch LBFGS (see this example).

Does your PR introduce any breaking changes?

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 list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or internal minor changes/refactorings)

PR review

Anyone in the community is welcome to review the PR.
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?

Make sure you had fun coding 🙃

[awaelchli]

Hi @lematt1991
Being able to attach a custom loop to the Trainer is an upcoming feature in 1.5, we call it "loop customization". Docs for it will be added soon #9557. To specify a custom FitLoop, you can simply set trainer.fit_loop = your_fit_loop directly on the trainer, or any of the subloops, for example trainer.fit_loop.connect(epoch_loop=...). We are not currently planning to add constructor arguments for these loops. Let me know if that helps.

Appreciate your initiative to send a PR!
In general, for features and API additions we would prefer to first receive an issue for discussion.