Trained on around ~170 images, suggested strength around 0.9-1.0. All images generated are using a-detailer.
