Welcome to my first style LoRA!
SD3 Version is already at Shakker AI!
https://www.shakker.ai/modelinfo/20d01bf7f9be44778b723939574346aa
Optimal settings and model to use it on:
Deliberate - https://civarchive.com/models/4823/deliberate
Anime - https://civarchive.com/models/108656/anime
Try to mix it with my "Eflorescense" LoRA! (See results in posts)
Prompt tweaks:
Optimal weights for LoRA: 0.7 - 0.9
"Trigger word" - Style by NTY. You can play with weights, I prefer to use it on 1.3 - 1.4.
Use "dark" for darker images, and "colourful" for more coloured arts.
Negative prompt:
(worst quality, low quality, mutated hands and fingers, bad anatomy, wrong anatomy, ugly, mutation:1.2)
Technical settings:
Steps: 22 - 27
Sampler: DPM++ 2M Karras
CFG scale: 2.5 - 5
Clip skip: 1 - 2 (preferably 2)
Feel free to play with weights to get the best results!
SDXL
I've made an SDXL version of my lora from the same dataset!
Description
FAQ
Comments (10)
gorgeous lora mia
Thank you, m.
Who is NTY? I want to see the original artist works.
NTY is a token unknown to Stable Diffusion used to call out LoRA without any premises.
Original author is Yuming Li on Twitter.
@OliviaOliveira Thanks for the reply, and this has been a fun lora to play with.
Love this LORA, I tried to make an embedding of this, but its no where as good as your LORA.
Would love to know what your training settings were, cos Im really interested in making Artstyle LORAs
You can view settings by tapping on "i" when hovering on LoRA in Stable Diffusion.
good
nice lora
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.