This LoRA was trained on images of the stained glass windows shot from inside the building. Expect bright glowy colours and many black lead lines as is traditional.

The training data was from pov camera shots from below, so it will bias that type of image.