Implementation for "Two Heads are Actually Better than One: Towards Better Adversarial Robustness via Transduction and Rejection" - View it on GitHub