ggml : unify tie-breaking to first index across all backends#25032
Draft
angt wants to merge 3 commits into
Draft
ggml : unify tie-breaking to first index across all backends#25032angt wants to merge 3 commits into
angt wants to merge 3 commits into
Conversation
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
Member
Author
|
not sure about CUDA yet 👀 |
github-actions
Bot
added
testing
Member
Author
|
well, Vulkan is not happy 👀 |
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
github-actions
Bot
added
SYCL
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Overview
This PR unifies
ggml_argmaxtie-breaking semantics across the CPU, Metal and WebGPU backends by returning the first occurrence of the maximum value, matching the behavior oftorch.argmaxandnumpy.argmax.It also adds regression tests covering tied maxima across different tensor sizes to ensure consistent behavior across backends.
Additional information
Found while experimenting with GLM-5.2 on the llama.cpp repository.
Requirements