Hi everyone! ✨
I trained this LoRA to capture the Kurukuru manga-style look I really love — soft but expressive linework, clean black-and-white panels, and that classic manga feeling with strong emotions and cute character expressions.
💖 Best Generation Settings
These settings usually give the most consistent results.
Aspect Ratio
- Best results: 1:1 (square images)
Recommended Settings
- Steps: 25–35
- CFG Scale: 5–7
- LoRA Weight: 0.7 – 0.9
Higher weights can push the style stronger, but may slightly overemphasize some learned features.
Example prompt structure:
"manga style, expressive character, dynamic pose, screentone shading, detailed ink lines"
🔞 NSFW
This LoRA can generate NSFW content depending on your prompts and the base model you use.
However, the training primarily focused on art style, so results may vary depending on the model and prompt structure.
🌸 Notes
- Works well for character-focused scenes and manga-style illustrations.
- Square compositions tend to keep the style balance and framing closer to the training images.
- Experimenting with LoRA weight and prompt wording can produce different stylistic variations.
Thank you for trying the model! I hope you have fun generating your own Kurukuru-style manga art ✨