Cpu memory accumulation bug #20730
Open
Cpu memory accumulation bug #20730
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ved1beta
requested review from
lantiga,
Borda,
tchaton,
justusschock and
ethanwharris
as code owners
seems this is not included in this PR yet |
for more information, see https://pre-commit.ci
…orch-lightning into cpu_memory_accumulation_bug
deependujha
reviewed
May 20, 2025
There was a problem hiding this comment.
Hi @ved1beta, thanks for the contribution.
Added some reviews. Please try fixing the failing tests and also mypy issue.
| # Clear memory if not returning predictions | ||
| import gc | ||
|
|
||
| gc.collect() |
There was a problem hiding this comment.
do you think it would be a good idea to have an argument collect_gc or something that users can toggle.
As Adrian said: it might be expensive in certain situations.
| if predictions is None: | ||
| self._warning_cache.warn("predict returned None if it was on purpose, ignore this warning...") | ||
| step_args = self._build_step_args_from_hook_kwargs(hook_kwargs, "predict_step") | ||
| step_output = call._call_lightning_module_hook(trainer, "predict_step", *step_args) |
There was a problem hiding this comment.
why are you directly calling lightning module hook without calling strategy hook?
After couple of checks and precision_plugin context, it does call lightning_module's predict_step.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This PR addresses the memory leak issue during prediction in PyTorch Lightning. It adds proper memory management when
return_predictions=Falseand includes comprehensive tests to verify the fix.Fixes #19398
Key Changes:
return_predictions=FalseBefore submitting
trainer.predict#19398PR 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
📚 Documentation preview 📚: https://pytorch-lightning--20730.org.readthedocs.build/en/20730/