Artist Style LoRA – Zaush
Artist: Zaush
LoRA Type: Artist Style
Trigger Word: z4ush
Recommended Weight: 0.9–1.2 (ideal around 1.15)
Base Model: Illustrious v1.1
Overview
This LoRA is designed to capture the signature artistic look of Zaush—a style defined cartoon-esque style with a bit of film grain and colored sketch styling.
Training Information
Steps / Epochs: 3348 steps across 15 epochs
Clip Skip: 2
Dataset Size: 372
Optimizer: Prodigy / Cosine
Training was performed on Illustrious v1.1, so models within the Illustrious ecosystem pair most accurately with this LoRA.
Recommended Generation Settings
To get results that match the intended style:
Sampler: Euler_a
Steps: 30
CFG: 4.5
Positive Prompt Suggestions
These optional tags help guide the aesthetic:
colored sketch, film grain, countershading, disney eyes (Used with the deer)Negative Prompt Suggestions
lowres, bad anatomy, bad hands, malformed hands, extra fingers, fused fingers, missing fingers, extra limbs, deformed, disfigured, multiple views, text, watermark, signature, username, copyright, error, jpeg artifacts, old, oldest, jaggy lines, latex, shiny skin, guroCompatibility
This LoRA works well with the following checkpoint:
WAI (Current Version): https://civarchive.com/models/827184/wai-illustrious-sdxl
Support the Artist
Please consider following or supporting Zaush.
This LoRA was created to help their style reach new audiences across the AI art community.
Respect the original artist and credit them whenever appropriate.
Have a question or suggestion?
I can usually be reached on Discord at: piper260
Description
FAQ
Comments (8)
Excellent stuff. These look super clean. What checkpoint/model are using?
EDIT:
Ah nevermind, I saw you mentioned WAI in the description. Thanks!
Absolutely!
Kk, so I'm quite sure that this model is the source of an oddity I'm experiencing. For whatever reason there is part of a human arm coming from offscreen and I have no idea what to do about it.
Feel free to DM me the prompt or workflow you're using and I can try to recreate it. I have also considered refining the dataset, for this model, soon. I just have a ton of images in this one. The grey mohawk hair is also an issue I've seen as well.
@Piper260 I'll try. The only other models I had were the character model, Eyes High Definition and Details Slider. Like 70-80% of the prompts are identical generation to generation. Only prompt I had was (((two hands, four fingers))) since without it the generated character tends to manifest a third hand where there shouldn't be one and the offscreen hand went away when I went back to a previous art style. I'm just stumped why this keeps happening, another art styles was causing ONE avian character to have human hair on their head when they don't have hair period.
@caelcyndar5559 What base model do you happen to be using? I've moved into using both Plantmilk - Walnut and WAI's v14 and I've been able to produce some solid results that didn't have any oddities in them.
@Piper260 Nova Furry XL. Was using WAI-NSFW-SDXL for all my Illustrious generations.
Any chance for an Anima version? The other Zaush LoRA doesn't quite get the style right.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.