Protecting Trained Models in Privacy-Preserving Federated Learning