]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
cuda : add warp-per-row wkv7 kernel for single-token decode (llama/26111)
author0 <redacted>
Tue, 11 Aug 2026 17:46:23 +0000 (01:46 +0800)
committerGeorgi Gerganov <redacted>
Fri, 14 Aug 2026 19:16:06 +0000 (22:16 +0300)
commit030656f26f139e7d94c92942d40424417046ca36
treeaf5d0ea26a4934a8ea6a6243f7102cf8064016ec
parentd9589688cbd4dd3a9e81c879cf72ff5a4eafcf40
cuda : add warp-per-row wkv7 kernel for single-token decode (llama/26111)
ggml/src/ggml-cuda/wkv.cu