Reset epoch progress with batch size scaler

[cschell]

What does this PR do?

1. This adds a test to check for the bug reported in scale_batch_size does not work anymore? #13696
2. Fixes the reported bugs. I'm not entirely convinced by my fix, it seems like this can easily break again: I fixed the issue by calling the reset functions of two deeply nested objects. Ideally, there should be a method dedicated to reset the state of the Trainer so that _run_binsearch_scaling and _run_power_scaling don't have to know about the Trainer internals. However, it seems to work fine for now, and since there is a test now for that particular issue, any breaking changes should at least get noticed.

Fixes #13696

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?

🦦

[codecov]

Codecov Report

Merging #13846 (dabf495) into master (6a999f1) will decrease coverage by 3%.
The diff coverage is 86%.

@@            Coverage Diff             @@
##           master   #13846      +/-   ##
==========================================
- Coverage      79%      76%      -3%     
==========================================
  Files         111      332     +221     
  Lines        7258    26909   +19651     
==========================================
+ Hits         5740    20450   +14710     
- Misses       1518     6459    +4941     

[rohitgr7]

will be fixed by #11089, but test could help :)