diff --git a/ggml/src/ggml-cuda/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu index 30ee84c5f4..3b079eb161 100644 --- a/ggml/src/ggml-cuda/ggml-cuda.cu +++ b/ggml/src/ggml-cuda/ggml-cuda.cu @@ -3903,6 +3903,7 @@ static void ggml_backend_cuda_graph_optimize(ggml_backend_t backend, ggml_cgraph const int root_node_idx = node_indices[root_node]; // only optimize for attn_norm + // TODO: make this generic if (!strstr(root_node->name, "attn_norm")) { continue; }