[Enhancement] Let optimizer use built torch opt #3642
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Outdated suggestions cannot be applied.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.
This PR slightly tweaks the constructor for
OptimWrapper
, by allowing it to use an already built pytorch optimizer. This is beneficial for when other libraries need to use a built optimizer directly to work with it. Prime need is Accelerate can prepare an optimizer, but it converts it back to pytorch. We then need to wrap it again inOptimWrapper
.Example use case:
To also have it be compatible with Accelerate,
step
takes in a closure param, but will raise an error if it's anything butNone
. This keeps it compatible with regular torch optimizers as well, though the implementation is not there (which @tmabraham is working on I believe)cc @jph00