You probably didn't expect this. As an experiment I trained this model on pictures of two heavily graffitied interpretive signs at a nearby nature walk. I'm testing multiple epochs because of the different effects the AI seems to have produced trying to figure out what it was seeing.