Fix checkpointed state for lr_schedulers with step interval

[simran2905]

What does this PR do?

Fixes #7637

The lr schedulers are being updated after calling on_train_batch_end and trainer._run_evaluation(), both of which can save checkpoints. This means that the latest state of lr_scheduler is not saved during checkpoint. This PR fixes the issue for classic schedulers.

NOTE: The ReduceLROnPlateau schedulers need on_train_batch_end called first for any metrics logging and will continue to save an older state potentially.

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 🙃

[pep8speaks]

Hello @simran2905! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-06-21 10:13:15 UTC

[codecov]

Codecov Report

Merging #7877 (bbbfce7) into master (2303f9c) will decrease coverage by 5%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #7877    +/-   ##
=======================================
- Coverage      92%     87%    -5%     
=======================================
  Files         210     210            
  Lines       13579   13581     +2     
=======================================
- Hits        12452   11812   -640     
- Misses       1127    1769   +642     

[tchaton]

Overall, looks good to me

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[awaelchli]

moved your changes from pytorch_lightning/trainer/training_loop.py to pytorch_lightning/loops/training_epoch_loop.py

[carmocca]

Pushed a few minor changes. Almost ready

[carmocca]

Changed the milestone to 1.4 as this requires several refactors not in the bug-fix branch

[simran2905]

Thanks @awaelchli and @carmocca for the merges.

[tchaton]

LGTM !

[tchaton]

Hey @simran2905,

quick question: Why differentiate reduce_on_plateau schedulers from the rest. Can't we call update_learning_rates only after callback metrics has been populated for all schedulers ?

Best,
T.C

[carmocca]

quick question: Why differentiate reduce_on_plateau schedulers from the rest. Can't we call update_learning_rates only after callback metrics has been populated for all schedulers ?

See the discussion in the linked issue: #7637 (comment)