This is a LoRA trained exclusively on a synthetic 3D character model created in character creation software (iClone / Character Creator). No real people were used in the creation or training of this model.

Model Name: NatalyX4 Trigger Word: NatalyX4

💡 Best Practices for Generating:

• Base Model: This LoRA was trained on SDXL Base and is highly optimized for photorealistic checkpoints like Juggernaut XL.

• Prompting: Use natural language (e.g., "NatalyX4, a photorealistic portrait of a woman...") instead of booru-style tags.

• Weights: Recommended LoRA weight is between 0.75 and 1.0. If the face looks "overbaked" or too harsh in your environment, lower the weight to 0.65 - 0.7.

⚙️ Technical Details:

• Designed to work seamlessly with ControlNet workflows (iClone to ComfyUI) for complex posing and lighting control.

• Highly flexible: The dataset was carefully captioned to separate the character's facial geometry from clothing, expressions, and environments.