chq2018515 / vib-pytorch

Pytorch implementation of Deep Variational Information Bottleneck

Deep Variational Information Bottleneck

Overview

Pytorch implementation of Deep Variational Information Bottleneck(paper, original code)

Dependencies

python 3.6.4
pytorch 0.3.1.post2
tensorboardX(optional)
tensorflow(optional)

Usage

1. train

python main.py --mode train --beta 1e-3 --tensorboard True --env_name [NAME]

2. test

python main.py --mode test --env_name [NAME] --load_ckpt best_acc.tar

References

1. Deep Learning and the Information Bottleneck Principle, Tishby et al.
2. Deep Variational Information Bottleneck, Alemi et al.
3. Tensorflow Demo : https://github.com/alexalemi/vib_demo