this project implements the Glow algorithm introduced in paper Glow: Generative Flow with Invertible 1x1 Convolutions
Glow is proposed to provide an invertible generative model. The project is implemented with utilities of tensorflow 2.0 and tensorflow probability.
pip3 install -U tfp-nightly tfp-nightly-gpu
pip3 install -U --pre tensorflow tensorflow-gpu
create dataset with the following command
python3 create_dataset.py
train model in eager execution mode with the following command
python3 train_eager.py