Variational Autoencoder models by tensorflow What is implemented? Variational Autoencoder which can embed high dimensional data such like image to low dimensional latent vectors which are supposed to be gaussian distribution.