Approximate Gradient Inversion Attack On Federated Learning