Fix auto scaling mode when calling tune method on trainer.

[ramonemiliani93]

What does this PR do?

Fixes #7319

Adding a test that should pass if an incorrect mode is used. The test also works to show that passing binsearch has no effect on the mode that is being run on the auto scaling. Not sure if the best way would be to update the scale_batch_size_kwargs in the tune method of the Trainer.

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?

Make sure you had fun coding 🙃

[codecov]

Codecov Report

Merging #7321 (76542c0) into master (e0c64f0) will decrease coverage by 7%.
The diff coverage is 100%.

@@           Coverage Diff            @@
##           master   #7321     +/-   ##
========================================
- Coverage      87%     81%     -7%     
========================================
  Files         200     200             
  Lines       12865   14380   +1515     
========================================
+ Hits        11210   11588    +378     
- Misses       1655    2792   +1137     

[awaelchli]

Found the issue. In the Tuner class line 46 we need to insert something like this:

            if isinstance(self.trainer.auto_scale_batch_size, str):
                scale_batch_size_kwargs.setdefault("mode", self.trainer.auto_scale_batch_size)

Thanks for adding the breaking test, that's perfect!!

[carmocca]

we need to insert something like this:

Exactly, sorry I broke this recently!

Pushed the fix

[SkafteNicki]

LGTM