Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix plateau scheduler stepping on incomplete epoch #8861

Merged
merged 7 commits into from
Aug 13, 2021

Conversation

awaelchli
Copy link
Contributor

@awaelchli awaelchli commented Aug 11, 2021

What does this PR do?

Fixes #8839

Added test FAILS without the fix applied.

This fixes an issue with the training loop trying to step an epoch level scheduler when the epoch is incomplete. This happens in the case when e.g. we specify Trainer(max_steps=x) where x < one epoch. This usually does not matter, but is problematic when the scheduler monitors a metric from a validation epoch, which does not actually run if we finish training early.

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?

I made sure I had fun coding 🙃

@awaelchli awaelchli added the bug Something isn't working label Aug 11, 2021
@awaelchli awaelchli added this to the v1.4.x milestone Aug 11, 2021
@codecov
Copy link

codecov bot commented Aug 11, 2021

Codecov Report

Merging #8861 (f08941e) into master (fec4f28) will decrease coverage by 4%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #8861    +/-   ##
=======================================
- Coverage      93%     89%    -4%     
=======================================
  Files         169     169            
  Lines       14078   14079     +1     
=======================================
- Hits        13052   12475   -577     
- Misses       1026    1604   +578     

@awaelchli awaelchli marked this pull request as ready for review August 11, 2021 23:07
tests/trainer/optimization/test_optimizers.py Outdated Show resolved Hide resolved
@mergify mergify bot added the ready PRs ready to be merged label Aug 12, 2021
@awaelchli awaelchli enabled auto-merge (squash) August 12, 2021 08:02
@awaelchli awaelchli merged commit 4b6aaee into master Aug 13, 2021
@awaelchli awaelchli deleted the bugfix/plateau-epoch branch August 13, 2021 01:35
four4fish pushed a commit to four4fish/pytorch-lightning that referenced this pull request Aug 16, 2021
kaushikb11 pushed a commit to kaushikb11/pytorch-lightning that referenced this pull request Aug 18, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working ready PRs ready to be merged
Projects
None yet
Development

Successfully merging this pull request may close these issues.

MisconfigurationException with ReduceLROnPlateau when taking a single step
4 participants