llama.cpp
vulkan: Use unclamped loads for flash attention mask
#12720

Merged

vulkan: Use unclamped loads for flash attention mask #12720

0cc4m merged 1 commit into ggml-org:master from jeffbolznv:flash_mask

jeffbolznv requested a review from

0cc4m 163 days ago

github-actions added testing

github-actions added Vulkan

github-actions added ggml

vulkan: Use unclamped loads for flash attention mask

12b198fa

jeffbolznv force pushed from 2df810a4 to 12b198fa 163 days ago

0cc4m approved these changes on 2025-04-06

0cc4m merged 80b717d4 into master 160 days ago

Reviewers

0cc4m

Assignees

No one assigned

Labels

testing Vulkan ggml

Milestone

No milestone