Feb 25, 2017 · 1 min read
That’s a very good question, I’ll talk about that in a following post about generative models! But basically it’s the same: the generator takes a random vector in the latent space (usually a vector of size 100) and creates an image from that.
