ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-03-08 21:10:06 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix DeepSeek FA (#382)

Browse Source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2025-05-05 08:39:10 +03:00
committed by GitHub
parent f7c9a0f036
commit e3fec17347
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ggml/src/ggml-cuda/fattn.cu
View File

@@ -521,6 +521,7 @@ void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst
// K and V head sizes. // K and V head sizes.
if (K->ne[0] != V->ne[0]) { if (K->ne[0] != V->ne[0]) {
ggml_cuda_flash_attn_ext_wmma_f16(ctx, dst); ggml_cuda_flash_attn_ext_wmma_f16(ctx, dst);
return;
} }
ggml_cuda_flash_attn_ext_mma_f16(ctx, dst); ggml_cuda_flash_attn_ext_mma_f16(ctx, dst);