The implement of paper "Minimum Adversarial Examples".
The calculation of the bound is based on "https://github.com/KaidiXu/auto_LiRPA"
First, you should install the 'auto_LiRPA'.
Then, download the dataset, we give an example of MNIST.
Last, run the 'test.py'