Skip to content

Commit a59f362

Browse files
committed
ggml-cpu: add comment for fattn disable
Signed-off-by: Aaron Teo <[email protected]>
1 parent b8e17f5 commit a59f362

File tree

1 file changed

+2
-0
lines changed

1 file changed

+2
-0
lines changed

ggml/src/ggml-cpu/ggml-cpu.c

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3526,6 +3526,8 @@ int ggml_cpu_has_sme(void) {
35263526

35273527
int ggml_cpu_support_fattn(void) {
35283528
#if defined(GGML_NNPA) || defined(__NNPA__)
3529+
// disable Flash Attention when using NNPA
3530+
// see: https://github.com/ggml-org/llama.cpp/issues/15721
35293531
return 0;
35303532
#else
35313533
return 1;

0 commit comments

Comments
 (0)