jramapuram / vae

Pytorch VAE's

VAE Pool

This repository is to be used as a submodule in projects where VAE's are required. It contains:

1. Sequentially reparameterized VAE (i.e. z0 --> z1 --> .. )
2. Parallelly reparameterized VAE (i.e. x --> [z0, z1] --> ..)
3. VRNN

All of the VAE implementations inherit from AbstractVAE.py which acts as a base abstraction.

Things to get a basic version working

You will need the helpers submodule from https://github.com/jramapuram/helpers.git This is needed as it provides basic functions such as utils and network builders. VAE's expect this helpers submodule to be at the baselevel of the repository.