This was trained using approx. 50 training sample images which had an unexpected quasi-Renaissance style that interested me. Of course I included some of the odder ones to liven things up.