Deponia style for Flux Reviews
I tested this lora several times and it has a lot of issues and none good or at least meh pictures.
0) Why only two preview pictures? Why is there none of metadata in LoRA? Because of that I can not understand how did you train it and adjust my prompts accordingly. Why description doesn't say anything about generating characters?
1) It often adds Rufus' beard to generated women. The result is simillar to Conchita Wurst (you can google this one, it's beardy man in women's clothing).
2) It often adds weird clothes which characters hold in hands (or try to put on?).
As I see, the issue lies in dataset. It's too small (how many pictures?) and it looks like there are not manypics with women only. Because of that it mixes women and men.
Another issue lies in not describing many concepts in original dataset, which became the style themselfes. Which, again, makes me think that dataset is very small (like 20 pictures?)