This is the second part to my first post on lipschitz constraining networks. Extending the idea to Wasserstein distances and modeling drift in systems.