🧪Inspired By Somi
Fine-tuned for use with SD 1.5-based checkpoints.
📌 Best Weight: 0.8
📌 CFG : 4.0
📸 Recommended Base Models:
supashymix / MakinaMix
🖼️ Training Resolution: 512x768 / 768x512 (optimized for portrait & vertical prompts)
✅ Tips for Best Results:
Use Hires.fix or ADetailer for improved sharpness & facial fidelity.
Works great for expressive and semi-stylized character renders.
📂 Available now on Ko-fi:
👉
💻 :
v1.0 :
- More realistic skintone.
- More detailed.
v2.0 :
- More defined headshape
- More defined facial features
Inspired By Somi v2.0
Discord : https://discord.gg/KsfFtUC794
Kofi : https://ko-fi.com/s/f3a3ceddc8
Patreon : d
✨ Support & Services:
Custom LoRA training available (DM or Ko-fi commission slot) Custom image generation requests accepted as well! Your support helps keep the models flowing ❤️
⚠️ This LoRA was not trained using celebrity images. Please refrain from generating likenesses of real individuals or public figures