[liuru]

Not at the moment in our tool, though this is an area of great curiosity and research for us!

[spywaregorilla]

Does the discriminator model translate the images into an embedding space of it's own? Could such a space be used to generate images themselves?