@Digital_Art_AI so yours is better in what way, if ? for example, I always test different loras and embeddings of the same person with a list of 96 different face expression ( prompt S/R) and sort those out which have the lowest face adaptability.

Because you'll find some of those other loras don't work on certain models very well. I've found Digital Art's lora's work a lot better on certain models where the others give me results that don't even look close.

@Digital_Art_AI What's wrong is you put no description or explanation on your model that does things differently or better than the other models. There is nothing wrong if you make something that is an improvement or has niche uses but how are we supposed to know if you don't write anything? If I were a new user I would just pick the one with most downloads already if there is nothing else that can guide me.

@Snipes Could you give examples and post pictures of the comparisons of the different LoRAs? I am genuinely curious. Though tbh this is OPs job.

@FaeFlan I don't have the time for that. Just fool around with different model and lora combos and you'll see they don't always work together. Sure it has something to do with what they were trained on. All I know is there are very few creator's celebrity loras that work on the current combo of model and other loras I've been using, but DAA's have been working. TI's rarely work well on the model's I use either, it's trial and error. So I welcome multiple versions of anything especially when they are trained on different models.