No idea how these people got their cats wedged into their neural nets
February 21, 2017 3:43 PM Subscribe
pix2pix-tensorflow uses machine learning to allow you to draw an outline and get back an image, constructed by mashing up pieces from a database of thousands of stock photos. In other words, you draw a outline of a cat, and you get back an image of a cat. Output ranges from the fairly straightforward to the horrifyingly twisted to the terrible mistake. The algorithm has also been turned loose on building facades, shoes, and handbags.
This thread has been archived and is closed to new comments