https://github.com/mwufi/pytorch-gqn

thinking of learning representations from video next, would be curious to see how that goes!