V1 trained with a small dataset of upscaled screenshots from Riven
Just did this for fun, because I always loved the game Riven.
