Make sharded checkpoint loading backwards-compatible #3240
Merged
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Outdated suggestions cannot be applied.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?
Loads metadata before loading sharded checkpoint state to check if 'optimizers' is a root-level key. This is for backwards compatability with older torch/composer versions, where the recommended sharded checkpoint loading/saving required us to put 'optimizers' as a top-level key. With newer torch (>=2.3.0), we don't do this anymore, but we still need checkpoints to be backwards compatible.
What issue(s) does this change relate to?
Before submitting
pre-commit
on your change? (see thepre-commit
section of prerequisites)