"Training Normalizing Flows with the Information Bottleneck for Competitive Generative Classification" (2020)

https://arxiv.org/abs/2001.06448

USAGE

• All configuration files for all experiments in the paper are contained in the directory 'experiments_configs'.

• Training (for one example configuration):

  python main.py train experiments_configs/cifar10/beta_ramp/beta_1.0000.ini

• Testing:

  python main.py test experiments_configs/cifar10/beta_ramp/beta_1.0000.ini`

• The cifar/mnist datasets should be downloaded automatically the first time it is run. For the OoD evaluation, tiny imagenet and quickdraw have to be downloaded separately.

REQUIREMENTS

To implement the INNs, we use of the FrEIA library (github.com/VLL-HD/FrEIA)

NOTE This code currently only works with the previous 0.2 version of FrEIA. To install it:

git clone https://github.com/VLL-HD/FrEIA.git
cd FrEIA
git checkout v0.2
python setup.py develop

Additional requirements:

Python >= 3.6

pip install -r requirements.txt