I've created a LoRA model for stable diffusion to depict scenes of Ultra Heroines defeating monsters, which was challenging to produce before.
This model specifically focuses on images where pink beams are striking the bellies of various monsters, including dinosaurs, demons, robots, and octopuses. Therefore, it's designed to generate scenes regardless of the location or type of monster. However, when using the default prompt 'giant monster', the model tends to produce mainly giant robot monsters. To get a wider variety, include specific prompts like 'dinosaur monster', 'demon', or 'octopus'.
Note that while the model is based on images where the beam hits the monster's belly in a straight line, the generated images might show scattered beams, so you may need to adjust the strength or regenerate multiple times to get a perfect shot.