Deep Directed Generative Models with Energy-Based Probability Estimation

code for the paper

Requirements

• Chainer 1.17

• PIL

• pylab

Contains the following repository:

• chainer-sequential

2D datasets

Train generator to generate 10 dimensional Gaussian mixture distribution and swiss-roll distribution.

See videos:

• https://gfycat.com/DarlingShowyHypsilophodon

• https://gfycat.com/UnrulyMisguidedHornedviper

Running

run train_2d/train.py to train the model.

run train_2d/gif_gaussian.py or train_2d/gif_swissroll.py to generate gif frames.

MNIST

run train_mnist/train.py

If there is no MNIST image, it will be downloaded automatically.

Genereted images

killmebaby

Download 686 images from http://killmebaby.tv/special_icon.html and resize all to 64x64 pixels.

run train_killmebaby/train.py

Original images

Images generated by Deep Generative Model

Since the position of the face of the training data is not constant, I think it is difficult to train the generator, but relatively clean images are generated.

When learning of Generator did not go well

Whichever noise z is used to generate an image, an average is generated.