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

Introduce Stateful DataModule #11637

Merged
merged 5 commits into from
Feb 7, 2022
Merged

Conversation

jjenniferdai
Copy link
Contributor

@jjenniferdai jjenniferdai commented Jan 26, 2022

What does this PR do?

part of / Fixes #11429: introduce Stateful DataModule (see Appendix code changes outlined in issue)

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

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

PR review

Anyone in the community is welcome to review the PR.
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 🙃

Copy link
Contributor

@ananthsub ananthsub left a comment

Choose a reason for hiding this comment

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

as a follow up, please deprecate CheckpointHooks off the DataModule interface: https://github.com/PyTorchLightning/pytorch-lightning/blob/86b177ebe5427725b35fde1a8808a7b59b8a277a/pytorch_lightning/core/datamodule.py#L26

It is no longer needed following the Stateful change and should be removed in v1.8

pytorch_lightning/core/datamodule.py Outdated Show resolved Hide resolved
tests/core/test_datamodules.py Outdated Show resolved Hide resolved
@ananthsub ananthsub added the checkpointing Related to checkpointing label Jan 28, 2022
@ananthsub ananthsub added this to the 1.6 milestone Jan 28, 2022
@jjenniferdai
Copy link
Contributor Author

jjenniferdai commented Feb 3, 2022

Did we want to deprecate CheckpointHooks off DataModule? The original issue #11429 discussion was more focused only on adding the Stateful alignment for all components (Motivation point 1) without editing CheckpointHooks (Motivation point 2) yet

@ananthsub
Copy link
Contributor

Did we want to deprecate CheckpointHooks off DataModule? The original issue #11429 discussion was more focused only on adding the Stateful alignment for all components (Motivation point 1) without editing CheckpointHooks (Motivation point 2) yet

I should have clarified that we can follow it up in a separate issue/PR. It is not a blocker for this one.

I initially proposed using the CheckpointHooks to DataModule back in #3544 . It was to capture dataloader state that had proprietary APIs. I wish at the time I had recognized the Stateful protocol lurking in pytorch and the rest of the lightning codebase as the contract for saving/loading state, but I didn't. I don't think there's a use case for the datamodule to be augmenting/changing other components state through the on_save/load_checkpoint hooks.

Copy link
Contributor

@ananthsub ananthsub left a comment

Choose a reason for hiding this comment

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

nice!

@ananthsub ananthsub added the lightningdatamodule pl.LightningDataModule label Feb 4, 2022
@mergify mergify bot added the ready PRs ready to be merged label Feb 5, 2022
@carmocca carmocca merged commit 1203094 into Lightning-AI:master Feb 7, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
checkpointing Related to checkpointing lightningdatamodule pl.LightningDataModule ready PRs ready to be merged
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[RFC] Standardize all stateful components on state_dict/load_state_dict
7 participants