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

WIP: Disable _optimizer_to_device logic #20036

Open
wants to merge 2 commits into
base: master
Choose a base branch
from

Conversation

corwinjoy
Copy link
Contributor

@corwinjoy corwinjoy commented Jul 2, 2024

This PR exists to test what happens when we disable _optimizer_to_device.

Pursuant to discussion in #19955

Before submitting
  • Was this discussed/agreed 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 list all the breaking changes introduced by this pull request?
  • Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist
  • 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

📚 Documentation preview 📚: https://pytorch-lightning--20036.org.readthedocs.build/en/20036/

@github-actions github-actions bot added the fabric lightning.fabric.Fabric label Jul 2, 2024
@corwinjoy
Copy link
Contributor Author

Running the tests locally this causes the following tests to fail:

> FAILED tests/tests_pytorch/loops/test_evaluation_loop_flow.py::test__eval_step__flow - RuntimeError: Expected all tensors to be on the same device, but found at l...
6712a7097
> FAILED tests/tests_pytorch/loops/test_training_loop_flow_scalar.py::test__training_step__epoch_end__flow_scalar - RuntimeError: Expected all tensors to be on the same device, but found at l...
6713a7099
> FAILED tests/tests_pytorch/trainer/test_trainer.py::test_multiple_trainer_constant_memory_allocated - AssertionError: assert device(type='cuda', index=0) == device(type='cpu')
6824c7210

In addition I would expect tests/tests_fabric/utilities/test_optimizer.py to fail where it checks all params are moved to the device. This test does not seem to run by default but I think we can agree that this test is incorrect since we want only some parameters moved to the device depending on the optimizer.

@awaelchli
Copy link
Contributor

Thanks for checking this @corwinjoy. I'll trigger all tests now in the CI because locally not all of them may run.

@awaelchli awaelchli changed the title Disable _optimizer_to_device logic WIP: Disable _optimizer_to_device logic Jul 2, 2024
@awaelchli awaelchli marked this pull request as ready for review July 2, 2024 10:57
Copy link

codecov bot commented Jul 2, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 57%. Comparing base (37e04d0) to head (d2589fc).
Report is 1 commits behind head on master.

❗ There is a different number of reports uploaded between BASE (37e04d0) and HEAD (d2589fc). Click for more details.

HEAD has 230 uploads less than BASE
Flag BASE (37e04d0) HEAD (d2589fc)
examples 6 0
cpu 91 36
pytest 59 21
app 6 0
gpu 4 3
lightning_fabric 13 7
lightning 58 26
lightning_app 5 0
pytorch_lightning 13 6
python3.11 28 14
python3.8 24 12
python3.10 22 10
pytorch2.1 24 12
pytest-full 36 18
pytorch2.3 6 3
pytorch2.2 6 3
Additional details and impacted files
@@            Coverage Diff            @@
##           master   #20036     +/-   ##
=========================================
- Coverage      83%      57%    -26%     
=========================================
  Files         427      422      -5     
  Lines       35460    35358    -102     
=========================================
- Hits        29599    20167   -9432     
- Misses       5861    15191   +9330     

@github-actions github-actions bot added the pl Generic label for PyTorch Lightning package label Jul 3, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
fabric lightning.fabric.Fabric pl Generic label for PyTorch Lightning package
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants