Commissions on Buymeacoffee or Ko-fi
🐦 Twitter | 🎨 Pixiv | 📷 Instagram
Some of you asked me to share my personal model. After some refinements and consistency tests, it finally here. This is huge merge that contains some checkpoints that I like:
All images samples were created with YOZORA.vae
This merge is capable of both male and female generations, but tends to work better with female characters. I tested this merge with a lot of character and style LoRAs, and the consistency is great, so don't mind burnt images, it never happened to me.
This is a no baked VAE version. So it works with most of them... But i recommend use VAEs that saturate images, i think it looks nicer and better.
For future versions, I'll try to bring a more cartoon look, but without throwing the 2.5d anime character that i like into the trash. It will be difficult, but we will try.
Feel free to use, enjoy or critisize.
This isn't a better version. It's just different
More faithful to the desired prompt
Tends to be more SFW
Less color contrast
Smoother skin
Loves freckles on the face (who doesn't?)
1. สิทธิ์ของโมเดลที่โพสต์ซ้ำเป็นของผู้สร้างต้นฉบับ
2. ผู้สร้างต้นฉบับที่ต้องการรับรองโมเดล โปรดติดต่อเจ้าหน้าที่ SeaArt AI ผ่านช่องทางทางการ คลิกเพื่อรับรอง
