This is an anime art style LoRA trained on SDXL / Illustrious.
This style LoRA is based on / inspired by the artwork of [SinderellaArt/Sina].
⚠️ Note: This is a Work in Progress (WIP). It's not 100% perfect yet, but I am actively tweaking the settings and dataset, and I will release updated versions over time! Feedback and test images are very welcome.
💡 Recommended Settings & Usage:
- LoRA Weight: 0.8 - 1.0
- Base Model: Illustrious / SDXL
- Trigger Word: Check each version below (each epoch has its own trigger word)
✍️ Prompting Guide:
- Prompt Structure: Put the LoRA and trigger word AT THE VERY BEGINNING of your positive prompt for maximum impact:
<lora:testsinderellaart-20:1.0>, testsinderellaart, 1girl, [your scene details...]
- Negative Prompt: Keep it simple or use standard quality negatives:
worst quality, low quality, bad hands, bad anatomy, missing fingers, extra digit, blurred
💡 Pro Tip for Experimenting:
You can also try using this LoRA WITHOUT the trigger word (just <lora:testsinderellaart-20:0.8>).
Depending on your prompt, omitting the trigger word can give a cleaner, less "overbaked" result with more natural details!
Description
Epoch 15 version
Best Results & Tested On:
- Trained on Illustrious base.
- Works best with Illustrious-based models (specifically tested on waiillustriousSDXL_v170).
- Recommended weight: 0.8 - 1.0.
- Recommended resolution: 1024x1536 (portrait) or 1536x1024 (landscape).
💡 Version Notes:
- Intermediate epoch checkpoint.
- May offer slightly more flexibility with different outfits, poses, and backgrounds compared to Epoch 20 if you find Epoch 20 too rigid.