Zaush, is a well-known furry artist who creating furry comics, illustrations, and other works. This is an attempt to capture his modern and most early art style, with two datasets of only pre 2010 and before images and Zaush's more modern works. To the time of Kama Citra and when he was mostly known for his full name Adam Wan. Enjoy!
Trigger word: Zaush
Anima Trigger word: @zaush
Recommended Anima Turbo LoRA Settings: Str-0.7 CFG-3-3.5 Steps-12
Description
The "Modern Day Zaush" LoRA ported over to Illustrious.
FAQ
Comments (15)
Heyo, which model are you using ? I am new to illustrious but see tons of models and would like one close to yours as a base
I'm a big fan of Nova Furry XL for image generation. https://civitai.com/models/503815/nova-furry-xl
Edit: I've updated the Suggested Resources with it too
If only it was that easy to make good friends irl like good pictures with this lora
How's everyone doing?
You can easily extract the config this model uses for training. You'll definitely find a guide online, or you can just ask some chatbot to find out
Behold the model, silent yet profound, where visions vast in quiet code are found. It bends no will, but gently lights the way, and births creations words can scarce convey. a muse in silicon, both wise and free— it shapes the world not as it is, but as it could be 🙃.
My dih went instantly limp when i saw on R34 what kind of age characters this guy drawing
"-We'll Handle these kids for you, Mr.Wan."
"-I think I know how to handle a kid, thank you."
Any plans for a Anima version?
I've got my eyes on it; however, I'm holding off testing anything on Anima until the final base model is released.
This style looks weird to me not bc anima
Let me know what you think of this latest version. I think it's much improved. But I'd appreciate the input.
@Murf_ New version is night and day different, might be 85%-ish there for the style. Still, you need pretty specific prompting to get the style going.
@OneGuyParty So, an "anchor tag" is a tag that when used with another tag will produce a specific effect. Like a style or a character. Unlike Illustrious, Anima really, really likes to group tags together and create anchors. Kind of trying to work out what tags in the dataset need to stay and what gets culled. Then there's the fact that Anima doesnt prompt like e621 so it's like learning a new language. A lora's Dataset always been 85% of the quality of a great trained lora, but it's even more important with Anima. No slacking.
Anyways, Thanks or the honest feedback <3
I appreciate you
@Murf_ That's pretty interesting. Also, using the Anima turbo lora, while it hurts the diversity of the gens, it boosts the style significantly. So there's definitely something going on with Anima, and not the dataset or the training.
Anyhow, thanks for the Lora!
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.