That’s pretty good, interesting that generative models seem to work better at representing the wilderness. Probably because it isn’t as regular as a civilized spaces