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

Seed all workers when using DDP #7942

Merged

Conversation

senarvi
Copy link
Contributor

@senarvi senarvi commented Jun 11, 2021

What does this PR do?

When seed_everything(workers=True) is called, it will set the environment variable PL_SEED_WORKERS=1. Consequently Trainer will set the worker_init_fn for dataloaders to pl_worker_init_function. However, this did not work with plugins that call reset_seed(), since that function calls seed_everything() with the default argument workers=False. This bug fix changes seed_everything() so that it will read the PL_SEED_WORKERS environment variable and pass the correct value for workers.

Fixes #7937

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 🙃

@senarvi
Copy link
Contributor Author

senarvi commented Jun 11, 2021

@awaelchli feel free to suggest better log messages.

@carmocca carmocca added bug Something isn't working distributed Generic distributed-related topic labels Jun 11, 2021
@carmocca carmocca added this to the v1.3.x milestone Jun 11, 2021
pytorch_lightning/utilities/seed.py Outdated Show resolved Hide resolved
pytorch_lightning/trainer/data_loading.py Outdated Show resolved Hide resolved
pytorch_lightning/utilities/seed.py Outdated Show resolved Hide resolved
@codecov
Copy link

codecov bot commented Jun 11, 2021

Codecov Report

Merging #7942 (ee0a16b) into master (20a5e09) will decrease coverage by 5%.
The diff coverage is 67%.

@@           Coverage Diff           @@
##           master   #7942    +/-   ##
=======================================
- Coverage      92%     87%    -5%     
=======================================
  Files         203     203            
  Lines       13129   13131     +2     
=======================================
- Hits        12017   11399   -618     
- Misses       1112    1732   +620     

@mergify mergify bot removed the has conflicts label Jun 13, 2021
@awaelchli awaelchli added the ready PRs ready to be merged label Jun 14, 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.

LGMT. Great catch !

@tchaton tchaton merged commit 22d8266 into Lightning-AI:master Jun 14, 2021
@carmocca carmocca mentioned this pull request Jun 15, 2021
carmocca added a commit that referenced this pull request Jun 15, 2021
* Seed all workers when using DDP

* Fix to dataloader seeding

* Make argument name explicit

Co-authored-by: Carlos Mocholí <[email protected]>

* Use f-strings when logging

* Removed a redundant log message

Co-authored-by: Carlos Mocholí <[email protected]>
lexierule pushed a commit that referenced this pull request Jun 17, 2021
* Seed all workers when using DDP

* Fix to dataloader seeding

* Make argument name explicit

Co-authored-by: Carlos Mocholí <[email protected]>

* Use f-strings when logging

* Removed a redundant log message

Co-authored-by: Carlos Mocholí <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working distributed Generic distributed-related topic ready PRs ready to be merged
Projects
None yet
Development

Successfully merging this pull request may close these issues.

worker_init_fn is not set to seed dataloaders correctly when using DDP
5 participants