Awesome

Auxiliary Deep Generative Models

Chainer implementation of the papers:

• Auxiliary Deep Generative Models
• Improving Stochastic Gradient Descent with Feedback
• Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks
• Categorical Reparameterization with Gumbel-Softmax

この記事で実装したコードです。

See also:

• VAE
• AAE

Requirements

• Chainer 1.17
• Pylab
• pandas

this code contains following repos:

• chainer-sequential

Runnning

ADGM

run train_adgm/train.py

SDGM

run train_sdgm/train.py

Validation Accuracy

Analogies

ADGM

SDGM