Add property to skip restoring optimizers and schedulers via plugin (DeepSpeed Checkpointing Fix 2/n) #8644
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Outdated suggestions cannot be applied.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?
Related to #8397
This PR adds a new property to skip optimizer/schedulers loading via the checkpoint connector, which for DeepSpeed is managed by the engine itself. The alternative was to move the loading logic into the plugin however this breaks the responsibility a bit for all other plugins that do not need this additional logic to be moved into the training type plugin.
Eventually, I'd like to see the actual loading logic end up inside the CheckpointPlugin, as described in #8118 to give the plugin a bit more control in the future.
CHANGELOG to be added in #8397
cc @awaelchli @tchaton
Before submitting
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:
Did you have fun?
Make sure you had fun coding 🙃