In the adversarial folder there is a set of tools to do attacks and robust training on neural networks.
In the verification folder there is one tool for neural network verification and one tool for decision tree verification.
In the relu_stability folder can be found the codes used to run experiment on ReLU stability and weight sparsity.
Please find in each folder more information about them.