Concept trained off of that one scene from Digimon Beatbreak where Sawashiro Kyo
tosses a minor arc villain off of a cliff. I thought the image was funny so I trained an entire concept off of it so you can have any other character toss another off of a cliff.
The prompt that seems to work the best is two views,the top view shows [character1] on the left y33t1ng [character2] on the right off a cliff,[character1 description],[character1] is standing with both arms forward and making a throwing motion, [character2 description] ,[character2] is falling back with arms slightly up, the bottom view is a close-up of [character2]'s face,[character2 facial expression]
Description
Base version of the model trained on a bunch of images I edited myself. Seems to work best with a strength of around 0.8. The trigger word is "y33t1ng", however you need to describe the scene for the model to produce it consistently. I recommend checking all of the example images, as they have the format you should use to trigger it optimally. It works with a wide variety of characters, including non-humans.