diff --git a/ggml/src/ggml-cuda/graph.cuh b/ggml/src/ggml-cuda/graph.cuh
new file mode 100644
index 00000000..ed032aa5
--- /dev/null
+++ b/ggml/src/ggml-cuda/graph.cuh
@@ -0,0 +1,41 @@
+#pragma once
+
+struct ggml_graph_node_properties {
+    void * node_address;
+    ggml_op node_op;
+    int64_t ne[GGML_MAX_DIMS];
+    size_t nb[GGML_MAX_DIMS];
+    void * src_address[GGML_MAX_SRC];
+    int32_t op_params[GGML_MAX_OP_PARAMS / sizeof(int32_t)];
+};
+
+struct ggml_cuda_graph {
+#ifdef USE_CUDA_GRAPH
+    ~ggml_cuda_graph() {
+        if (instance != nullptr) {
+            CUDA_CHECK(cudaGraphExecDestroy(instance));
+        }
+        if (graph != nullptr) {
+            CUDA_CHECK(cudaGraphDestroy(graph));
+        }
+    }
+    cudaGraph_t graph = nullptr;
+    cudaGraphExec_t instance = nullptr;
+    size_t num_nodes = 0;
+    std::vector<cudaGraphNode_t> nodes;
+    std::vector<cudaKernelNodeParams> params;
+    bool disable_due_to_gpu_arch = false;
+    bool disable_due_to_too_many_updates = false;
+    bool disable_due_to_failed_graph_capture = false;
+    int number_consecutive_updates = 0;
+    std::vector<ggml_graph_node_properties> ggml_graph_properties;
+    bool use_cpy_indirection = false;
+    std::vector<char *> cpy_dest_ptrs;
+    char ** dest_ptrs_d;
+    int dest_ptrs_size = 0;
+    // Index to allow each cpy kernel to be aware of it's position within the graph
+    // relative to other cpy nodes.
+    int graph_cpynode_index = -1;
+#endif
+};
+