]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
ggml-cpu/ops: vectorize flash-attention V-cache F16 to F32 conversion (llama/26947)
authorjinzihao <redacted>
Thu, 13 Aug 2026 09:04:46 +0000 (17:04 +0800)
committerGeorgi Gerganov <redacted>
Fri, 14 Aug 2026 19:16:06 +0000 (22:16 +0300)
commit62031fed0b4a2f16ba4fc3bb4567cfca19d89dfa
tree7a92be21908d39f5a8477927fbe3e6feff0566ae
parentac9a74f1a993bfd086666689aec7c3d16920ce6b
ggml-cpu/ops: vectorize flash-attention V-cache F16 to F32 conversion (llama/26947)

Co-authored-by: jinzihao <redacted>
ggml/src/ggml-cpu/ops.cpp