🌸 Virtual Japanese Woman RIAN
💡 Note
In my environment, when HighNoise and LowNoise were trained separately, the face turned out slightly different.
Therefore, as with Wan2.1, please use a single merged LoRA, applied at 1.0 for both HighNoise and LowNoise, when generating videos or images.
If you don’t use both, it hasn’t been thoroughly tested, but LowNoise alone should work fine.
Regarding the face, I’ve made substantial adjustments; however, overseas users—and even people from East Asia—may perceive them as only subtle differences. Thank you for your understanding.
Trained on 1024x1536 images to reproduce a photorealistic Japanese woman style named RIAN.
Not a glamorous idol, but rather the vibe of a “cute woman you might encounter in everyday Japan.”
🔹 Trigger word: rian (likely works even without the trigger)
🔹 Recommended weight: 0.7 – 1.0
⚠️ This character is fictional and does not exist in reality.
⚠️ Facial appearance may vary depending on other LoRAs or workflows used.
⚠️ Highly prompt-dependent: please refine features via prompt (e.g., bangs, blush, lip gloss, gaze, angle, expressions).
🟢 Limited testing showed good results. As it’s still in an experimental phase, generation may occasionally fail.