Code for reproducing key results in the paper InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets by Xi Chen, Yan Duan, Rein Houthooft, John Schulman, Ilya Sutskever, Pieter Abbeel.
Line or Rectangles datasets are available on s3.
# Single Line. Foreground Noise. (14.5 mb)
curl https://whiskey-ginger-analytics-public.s3.amazonaws.com/datasets/BASICPROP.zip | tar -xf- -C ./
# Pair of Rectangles. (579 kb)
curl https://whiskey-ginger-analytics-public.s3.amazonaws.com/datasets/BASICPROP-angle.zip | tar -xf- -C ./
# Pair of Rectangles. Foreground Noise. (53.7 mb)
curl https://whiskey-ginger-analytics-public.s3.amazonaws.com/datasets/BASICPROP-angle-noise.zip | tar -xf- -C ./
# Pair of Rectangles. FG + Background Noise. (76.1 mb)
curl https://whiskey-ginger-analytics-public.s3.amazonaws.com/datasets/BASICPROP-angle-noise-bg.zip | tar -xf- -C ./
This project currently requires an antiquated version of tensorflow. For Mac:
pip install -U https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.10.0-py2-none-any.whl
In addition, please pip install the following packages:
prettytensorprogressbarpython-dateutil
$ git clone [email protected]:openai/InfoGAN.git
$ docker run -v $(pwd)/InfoGAN:/InfoGAN -w /InfoGAN -it -p 8888:8888 gcr.io/tensorflow/tensorflow:r0.9rc0-devel
root@X:/InfoGAN# pip install -r requirements.txt
root@X:/InfoGAN# python launchers/run_mnist_exp.pyWe provide the source code to run the MNIST example:
PYTHONPATH='.' python launchers/run_mnist_exp.pyYou can launch TensorBoard to view the generated images:
tensorboard --logdir logs/mnistMIT