👉 MoreTwitch Streamers / OnlyFans LoRAs HERE👈
A weight of 1 is fine.
The pictures are a workaround for Civitai TOS. The pictures you get from the LoRa are usually much closer to what she usually wears.
🙏 If you like this model, please click the Like button and leave a 5 stars review, it really helps for visibility! 🙏
You can now request a custom LoRa on 👉 Fiverr 👈 or on 👉 buymeacoffee (no charges)👈
Description
FAQ
Comments (3)
never before tried to load a human-based lora on an anime-checkpoint. thought i get some anime-looking ariana, instead i got a plain, generic anime girl. pretty disappointed to be honest.
lesson learned. will never do that again :(
EDIT: and you are right with tos.. even anime gets its tos here. well, let's hope, the other 3 images get approved.
EDIT2: loading this lora together with "beautyfool" generated EXACTLY the images that i had in mind. now let's hope for an approval for this one, too.
Thank you for your comment. Most of my loras work well on both anime and realistic. My LoRas are not overtrained so you can still higher the weight of lora and trigger word to more than 1, and cfg up to 9 or 10. It should be strong enough to affect anime models. You can also use some 2.5D model. I added some anime pictures so you can see the settings.
I have to add that Ariana is very beautiul but she has a very generic body and face, which makes it more difficult to recognize her as anime form.
Transforming people into anime is something very funny to do so i strongly recommend you to continue to experiment with it, even if this lora is not the best for this.
Have a good day.
I feel like newer models are more resistant to loras in general, maybe the more they get mixed, and loras in the mixes as well, is making it harder for models to integrate additional conditioning. I can hardly get them to follow prompts either, but I feel like it's some settings on my end. Been messing with earlier versions of older models, and a bit surprised how good they can be vs. recent versions.