And... Another character LoRA, Diana!

I had a lot of fun creating and testing her. As always, feedback, images and suggestions are very welcome — I'd love to see what you create with her! ❤️

Diana is a photorealistic female character with a soft and elegant appearance, subtle freckles, light eyes and naturally feminine features.
The LoRA was trained to preserve her facial identity and natural proportions while remaining flexible with hairstyles, clothing, poses, environments and different visual styles.

Recommended weights

• 0.9 – 1.0 → Best for a very faithful Diana, close to the original character and training dataset.

• 0.75 – 0.9 → Recommended for more freedom with hairstyles, outfits, makeup and creative variations while keeping Diana recognizable.

• Below 0.75 → Useful when you only want some of Diana's facial characteristics and more influence from the base model.

Notes

• Trigger word: Diana
• Hairstyles, hair color, eye color, makeup and clothing can be changed.
• Works well with portraits, full-body images and different camera angles.
• Diana's identity remains surprisingly consistent even with significant hairstyle and makeup changes.
• For maximum likeness, I recommend using a weight around 0.9–1.0. (I'm using 0.92)

Have fun with Diana! ❤️