This repo contains code of our term project for Machine Learning course in Peking University.
File structure: 2 directories, 14 files
├── Lipschitz : Lp-Network components
│ ├── core.py : Functions used to calculate Lp-Neuron output
│ └── nn.py : Lp-Neuron implementation of some torch.nn layers
│
├── mnist.py : Experiment on MNIST to investigate expressive power
├── robust.py : Experiment on MNIST to compare certified robustness
├── plot.ipynb : Visualization scripts
│
└── output : Output of our experiments
├── ...
Reproduce results:
- python mnist.py <p>
- python robust.py