This model aims to offer realistic but, flexible styles. The model was trained with over 800 images. If you want the dataset just message me.