Fast generative models composed with neural networks, powered by PyTorch
This is used for personal research so the code is far from bug-free.
Implementations of various variants of Variational Autoencoders:
- Vanilla VAE:
Diederik P. Kingma, Max Welling. Auto-Encoding Variational Bayes. 2013
- Gumbel-softmax trick to backprop through discrete variables:
Eric Jang, Shixiang Gu, Ben Poole. Categorial Reparameterization with Gumbel-Softmax. 2017
- Semi-supervised VAE:
Diederik P. Kingma, Danilo J. Rezende, Shakir Mohamed, Max Welling. Semi-supervised Learning with Deep Generative Models. 2014
- VAE using auxilary variables:
Lars Maaløe, Casper Kaae Sønderby, Søren Kaae Sønderby, Ole Winther. Auxiliary Deep Generative Models. 2016
- Distangled representations using β-VAE:
Irina Higgins, Loic Matthey, Arka Pal, Christopher Burgess, Xavier Glorot, Matthew Botvinick, Shakir Mohamed, Alexander Lerchner. β-VAE: Learning basic visual concepts with a constrained variational framework. 2017
- Balancing Learning and Inference in VAE:
Shengjia Zhao, Jiaming Song, Stefano Ermon. InfoVAE: Information Maximizing Variational Autoencoders. 2018