This model produces some silenthill-esque abstract subjects and landscapes.

The pony version kinda works, kinda doesnt, more portraits than environmental

Il be working on retraining and refactoring it.