The model I think is fairly self explanatory.

So with this one, I don't know. It works for some things, some things not so much, the request was for Dragon Ball Super stages, such as Zeno's palace, and that -is- in the dataset, but the only thing it really does well is kind of natural environments as you can see with the previews. Typically I don't like to add previews of images that 'don't work', but I'll do it here because it's rather bizarre and I find it funny how much it didn't work. The 'tags' I'll add to the trigger sections are the ones I feel like work to generate -something- in the room of something you might want. I'll fully admit that I'm kind of awkward at doing backgrounds for LoRA's, but I'm sort of learning them as I go along since I've never really played with them.

Works well at a strength of 1.

Previews made with novaFurryXL_illustriousV6.b