Fix logging on_train_batch_end in a callback with multiple optimizers

[carmocca]

What does this PR do?

Fixes #5459

High-level explanation:
Before, we did range(num_optimizers) but this fails because epoch_metrics can contain data for just one optimizer.

For this reason, the fix is to loop using the epoch_metrics keys. list(epoch_metrics) (to use a copy of the keys) is necessary because keys are deleted inside the loop.

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?
• [n/a] 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

• 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
• Check that target branch and milestone match!

[codecov]

Codecov Report

Merging #5521 (6ffc2b0) into master (a56f745) will decrease coverage by 0%.
The diff coverage is 100%.

@@          Coverage Diff           @@
##           master   #5521   +/-   ##
======================================
- Coverage      93%     93%   -0%     
======================================
  Files         135     135           
  Lines       10007   10005    -2     
======================================
- Hits         9341    9339    -2     
  Misses        666     666           

[tchaton]

LGTM !

[ananthsub]

thanks for the fix!