Edges-to-cats model trained with 700 images x 200 epochs (from...