Add support for automatic mixed precision #31
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The tests pass, but they shouldn't -- torch.amp was only added in 1.10. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 pull request adds decorators to
EntmaxBisectFunction,SparsemaxBisectFunction,Entmax15Function, andSparsemaxFunctionso that they will autocast to 32-precision if they occur inside an autocast context. This seems to be the right thing to do because bf16 and fp16 introduce numerical stability issues that cannot easily be solved. Upcasting should make it easy to incorporate entmax into any mixed-precision setting.The actual code that needed to be written was incredibly simple; the vast majority of commits are me making silly mistakes while writing the tests.