✨ Like the results?
Download this Checkpoint for unlimited local use and support future creations, all versionin one.
☕ Unlock the download on Ko-fi →https://ko-fi.com/s/f3f62073b4
Credtis showcase:
Image 3 prompt by Zumu (https://tensor.art/u/792127093621175955)
🧪 AII-KO Illustrious
A fine-tuned model without a fixed base style, making it ideal for adapting with LoRAs or experimenting with different artist styles without being limited to a single predefined aesthetic.
Base Settings:
Quality Prompt: masterpiece, good quality, absurdres, highres
Negative Prompt: lowres, bad quality, worst quality, bad , jpeg artifacts, ugly, poorly drawn, censor, watermark, text, bad fingers, deformed fingers
Resolution: Up to 1024×1536 Sampler: Euler A Steps: 25–42 CFG Scale: 3–7
Calling Characters Recognized by the Model
Use the following format:
character name \(series/game/anime\)
Example:
ruan mei \(honkai: star rail\)
Character List: You can browse the available character tags on Danbooru.
Prompt Guide: You can use the Danbooru Wiki as a prompt reference.
Prompt Style: Booru tags
Suggested Starting Prompt:
masterpiece, good quality, absurdres, highres, anime screencap, 1girl, solo, yuinshiel Asteria, blonde hair, blue background, blue eyes, blush,, long hair, looking at viewer, on throne, open mouth, simple background, sitting, skirt, thighhighs, throne, uniform, white thighhighs
Having Issues with Hands?
Try this UNet-only hand-fixing LoRA:
https://tensor.art/models/898461132648391071
Other LoRAs Trained on This Model:
AII-KO Anime Style
https://tensor.art/models/845825712511487326
748cm Style
https://tensor.art/models/844695049485836850/748cm-A-2B-E22
Anime Style for Illustrious
https://tensor.art/models/805553462034198523/Anime-Style-for-Illustrious-A_Style2B
Honkai Star Rail Style – AIIKO
https://tensor.art/models/1003055283649383386/Honkai-Star-Rail-Style-AIIKO-v1.0-Anima-Unet
Complete Library:
AIIKO Models - Best AI Art & Models | (https://tensor.art/u/622954954671366110/models)Tensor.Art (https://tensor.art/) | (https://tensor.art/u/622954954671366110/models)Tensor.Art (https://tensor.art/)
Changelog:
v1. First optimal release
v2. Improvements to the adaptive base style and some fixes.
v3.2. Final optimization adding new knowledge, style adjustments, and prompt.
Which is optimal?
It will depend on what you find best from your point of view. In my opinion, each version excels in something, whether it's an artistic style, prompt, etc.
Want to contact me?
Discord: hirayama.ai