]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
CUDA: extend K-type validation to V-types for flash attention (llama/24403)
authorAlexey Kopytko <redacted>
Mon, 6 Jul 2026 14:26:50 +0000 (23:26 +0900)
committerGeorgi Gerganov <redacted>
Fri, 10 Jul 2026 10:06:42 +0000 (13:06 +0300)
commit82ac928a5d832f0e7d56c5896aa51269b78b2e2f
treef8338a205764a05f02d836d3adc8dddd20898a92
parent7d8e64fe373ea9a0da7230e6b01f9ea3879a5dc1
CUDA: extend K-type validation to V-types for flash attention (llama/24403)

* CUDA: extend K-type validation to V-types for flash attention

* reorder
ggml/src/ggml-cuda/fattn.cu