Skip to content

Conversation

@ananthsub
Copy link
Contributor

@ananthsub ananthsub commented Aug 23, 2021

What does this PR do?

  • Moves the validating logic of checking for the path's existence from the checkpoint connector to the torch checkpoint IO plugin (where the fsspec logic lives)
  • Moves the CUDA cache clearing from the checkpoint connector to the training type plugin
  • Renames load_checkpoint_file in the training type to load_checkpoint for consistency with the checkpoint IO's API

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

This is a breaking change for any custom plugin that extended load_checkpoint_file as this has been renamed to load_checkpoint with this PR

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 🙃

@codecov
Copy link

codecov bot commented Aug 23, 2021

Codecov Report

Merging #9045 (dbb65b9) into master (8a93173) will decrease coverage by 4%.
The diff coverage is 91%.

@@           Coverage Diff           @@
##           master   #9045    +/-   ##
=======================================
- Coverage      93%     89%    -4%     
=======================================
  Files         175     175            
  Lines       14379   14457    +78     
=======================================
- Hits        13322   12806   -516     
- Misses       1057    1651   +594     

@awaelchli awaelchli added this to the v1.5 milestone Aug 23, 2021
@mergify mergify bot added the ready PRs ready to be merged label Aug 23, 2021
Co-authored-by: Adrian Wälchli <aedu.waelchli@gmail.com>
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.

Neat !

@ananthsub ananthsub merged commit 1e4d892 into Lightning-AI:master Aug 23, 2021
@ananthsub ananthsub deleted the fix/rename-load-checkpoint-plugin branch August 23, 2021 20:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

breaking change Includes a breaking change checkpointing Related to checkpointing ready PRs ready to be merged refactor

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants