Conditional Variational Autoencoder (CVAE)

References

• Kingma et al., "Semi-supervised learning with deep generative models," NIPS 2014
• Sohn et al., "Learning structured output representation using deep conditional generative models," NIPS 2015

What I've done

• Conditioned on MNIST images, we want to generate images in the different labels. In this code, for example, if the image with label 1 is given as condition, we want to generate the image with its label 0 or 2 with probability one half.