Update "phantom grad" to use the same forward-backward transformation code as "vjp" #364
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
IvanYashchuk
requested review from
mruberry,
lantiga,
robieta,
t-vi and
carmocca
as code owners
IvanYashchuk
commented
May 13, 2024
|
All checks have passed. @mruberry, could you please help merge this PR if the changes make sense? |
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Outdated suggestions cannot be applied.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 PR swaps
thunder.transforms.gradwiththunder.transforms.grad_v1. The difference between the two is that grad_v1 uses the same code as the main code path with PyTorch Autograd integration, they differ only in the implementation but do the same thing.After this PR, I'm planning to remove the
gradfunction code that is now unused and renamed in this PR to__grad.The grad transform was added in 6069e6f and it included a new way of defining forward-backward rules in one same Python function and a new way of registering a transformation to the jitted function.
The new way of defining forward-backward rules is now also supported in the main code path; we no longer need that prototype.
Test plan:
pytest thunder/tests/test_grad.py -k "phantom_grad".