V8
Best style, best concept tags. I recommend this one.
slightly different from v7.
V7
Without any trigger words(sample of ganyu), it makes great styles.
V6
Some better tags might help (like 'nipple markings' if you want to avoid text writings on the body). But for some reason, the style looks a bit different—the outlines appear thicker.
I'd prefer V5 style more, but V6 has better concept expressions.
V5
trained on new datasets.
4 style trigger words: prerbqinori, midrbqinori, postrbqinori, recrbqinori,
Description
FAQ
Comments (4)
It's amazing, this is the king of this kind of lora
amazing lora, lots of concepts baked in for people to try out.
Will you train a pure concept lora without artist style? The current one seems respond too much to the artist style. But still great work!
This is a high quality Style Lora - it really delivers on those prompts - great job!