I wrote a DIVA (DIVergent Autoencoder) model for the Catlearn R library

Check it out here.