For now, we show that if StarGAN is trained on images augmented by adversarial attack it does become more resistant but not completely resistant to attacks.