PyTorch implementation of the variational recurrent neural networks (VRNN).

• Chung, Junyoung, Kyle Kastner, Laurent Dinh, Kratarth Goel, Aaron C. Courville, and Yoshua Bengio. A recurrent latent variable model for sequential data. NeurIPS, 2015.

This model is also known as the recurrent state space model (RSSM).

• Hafner, Danijar, Timothy Lillicrap, Ian Fischer, Ruben Villegas, David Ha, Honglak Lee, and James Davidson. Learning latent dynamics for planning from pixels. PMLR, 2019.

A version with discrete latent space and Gumbel Softmax reparametrisation is implemented in branch discrete.

See results in the Weights and Biases repository.