Repository navigation
Preserve input dtype in FuseBiasInLinear - #2533
Open
gorkadabo1 wants to merge 1 commit into
Open
gorkadabo1 wants to merge 1 commit into
gorkadabo1 wants to merge 1 commit into
Conversation
3 tasks done
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
Fixes #2532
FuseBiasInLinearcreates the appended ones tensor with the default dtype, which promotesfloat16andbfloat16inputs tofloat32and causes a dtype mismatch with the fused weights.This change passes
linear_input.dtypeto the existingtorch_onescall, preserving the input dtype. It also adds five focused regression cases coveringfloat32,float16,bfloat16, and conversion to reduced precision after transformation.Validation
float32case passed, while the other 4 failed with the expected dtype mismatch.git diff --checkpassed.Tests were run on Hyperion with PyTorch 2.14.1 and Transformers 4.57.6.
Before submitting
Who can review?
@IlyasMoutawwakil