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 retrieval of batch indices when dataloader num_workers > 0 #10870

Merged
merged 21 commits into from
Dec 2, 2021

Conversation

awaelchli
Copy link
Contributor

@awaelchli awaelchli commented Dec 1, 2021

What does this PR do?

Fixes #10782

Please see my comment here #10782 (comment) for a detailed answer where the bug comes from and why we need to change the way we use the IndexBatchSamplerWrapper. In short, the way a PyTorch BatchSampler retrieves the batch indices differs between num_workers=0 and num_workers > 0 in the DataLoader, and our prediction loop needs to account for that.

Does your PR introduce any breaking changes? If yes, please list them.

No.

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 🙃

Part of #1 (it's a lie, this is just here to avoid noisy GitHub bot)

cc @Borda

@awaelchli awaelchli added bug Something isn't working callback: prediction writer labels Dec 1, 2021
@awaelchli awaelchli added this to the 1.5.x milestone Dec 1, 2021
Copy link
Contributor

@tchaton tchaton left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM !

@mergify mergify bot added the ready PRs ready to be merged label Dec 2, 2021
pytorch_lightning/overrides/distributed.py Outdated Show resolved Hide resolved
pytorch_lightning/overrides/distributed.py Outdated Show resolved Hide resolved
tests/deprecated_api/test_remove_1-7.py Outdated Show resolved Hide resolved
@mergify mergify bot added the has conflicts label Dec 2, 2021
@mergify mergify bot removed the has conflicts label Dec 2, 2021
Co-authored-by: Rohit Gupta <rohitgr1998@gmail.com>
@awaelchli awaelchli enabled auto-merge (squash) December 2, 2021 09:42
@codecov
Copy link

codecov bot commented Dec 2, 2021

Codecov Report

Merging #10870 (ed735c0) into master (541b983) will increase coverage by 0%.
The diff coverage is 100%.

@@           Coverage Diff           @@
##           master   #10870   +/-   ##
=======================================
  Coverage      92%      92%           
=======================================
  Files         177      177           
  Lines       16567    16579   +12     
=======================================
+ Hits        15206    15218   +12     
  Misses       1361     1361           

@awaelchli awaelchli merged commit c55bc43 into master Dec 2, 2021
@awaelchli awaelchli deleted the bugfix/batch-indices branch December 2, 2021 10:36
awaelchli added a commit that referenced this pull request Dec 2, 2021
Co-authored-by: Rohit Gupta <rohitgr1998@gmail.com>
lexierule pushed a commit that referenced this pull request Dec 7, 2021
Co-authored-by: Rohit Gupta <rohitgr1998@gmail.com>
@rohitgr7 rohitgr7 mentioned this pull request Feb 7, 2022
12 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working callback: prediction writer ready PRs ready to be merged
Projects
None yet
Development

Successfully merging this pull request may close these issues.

batch_indices passed to PredictionWriter write_on_epoch_end is wrong when strategy=ddp
4 participants