fix: compare inputs_embeds generate against input_ids generate, not u…#1787
Draft
fix: compare inputs_embeds generate against input_ids generate, not u…#1787
Conversation
…ncached decode The test compared cached generation (model.generate) against uncached manual decoding (use_cache=False). For Mamba hybrid models, these two paths use different CUDA kernels that are mathematically equivalent but not bit-identical in bf16, causing spurious assertion failures. Changed the reference to use generate(input_ids=...) which exercises the same cached kernel path as generate(inputs_embeds=...). Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 2897aab |
Signed-off-by: Alexandros Koumparoulis <akoumparouli@nvidia.com>
Contributor
Author
|
/ok to test 50f5374 |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
…ncached decode
The test compared cached generation (model.generate) against uncached manual decoding (use_cache=False). For Mamba hybrid models, these two paths use different CUDA kernels that are mathematically equivalent but not bit-identical in bf16, causing spurious assertion failures.
Changed the reference to use generate(input_ids=...) which exercises the same cached kernel path as generate(inputs_embeds=...).
What does this PR do ?
Add a one line overview of what this PR aims to accomplish.
Changelog
Before your PR is "Ready for review"
Pre checks:
If you haven't finished some of the above items you can still open "Draft" PR.
Additional Information