From e079bffb6678e1b5ded21c719600bedd7175e726 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Sun, 8 Sep 2024 22:01:02 +0300 Subject: [PATCH] cuda : fix FA Q src index (1 -> 0) (#9374) --- ggml/src/ggml-cuda/fattn.cu | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/ggml/src/ggml-cuda/fattn.cu b/ggml/src/ggml-cuda/fattn.cu index f87f33b3e574b..f28a19d40b356 100644 --- 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];