graph : fix graph reuse reset of params

ggml-ci

src/llama-graph.cpp

@@ -428,6 +428,8 @@ void llm_graph_result::reset() {
     t_embd        = nullptr;
     t_embd_pooled = nullptr;
 
+    params = {};
+
     inputs.clear();
 
     buf_compute_meta.resize(ggml_tensor_overhead()*max_nodes + ggml_graph_overhead_custom(max_nodes, false));