First-Order Adversarial Vulnerability of Neural Networks and Input Dimension - View it on GitHub