StyleCLIP, manipulating StyleGAN imagery with text

Where do we go…


Input: Photo of a politician
Output: politician with a soul and some grasp of ethics


whoah there phil… that’s not possible with current technology.

absolutely savage. brutal.

Does that fall under the pick two paradigm?

Or the pick one paradigm?


That’s the faded, crumbly story of my culturally impoverished existence Bryan.

“Oh the next version has the super tool that you want, it will be ready tomorrow…”

“P.S. it won’t run on your current hardware…”