cuda : fix FA Q src index (1 -> 0) (ggerganov#9374)

CISC · Sep 8, 2024 · e079bff · e079bff
1 parent 3f7ccfd
commit e079bff
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/ggml/src/ggml-cuda/fattn.cu b/ggml/src/ggml-cuda/fattn.cu
@@ -152,7 +152,7 @@ static void ggml_cuda_flash_attn_ext_wmma_f16(ggml_backend_cuda_context & ctx, g
     }                                                                       \
 
 static void ggml_cuda_flash_attn_ext_vec_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
-    ggml_tensor * Q = dst->src[1];
+    ggml_tensor * Q = dst->src[0];
     ggml_tensor * K = dst->src[1];
     ggml_tensor * V = dst->src[2];
 
@@ -227,7 +227,7 @@ static void ggml_cuda_flash_attn_ext_vec_f16(ggml_backend_cuda_context & ctx, gg
     }                                                                       \
 
 static void ggml_cuda_flash_attn_ext_vec_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
-    ggml_tensor * Q = dst->src[1];
+    ggml_tensor * Q = dst->src[0];
     ggml_tensor * K = dst->src[1];
     ggml_tensor * V = dst->src[2];