AvaviLora is a character and style LoRA specialized in generating photorealistic, film-quality Na'vi characters. Trained on highly detailed , natural language captions, this model excels at capturing the unique physiology, intricate accessories, and atmospheric lighting of Pandora.

Unlike simple character models, AvaviLora understands complex facial expressions, dynamic poses, and emotional storytelling.

ℹ️ Model Details

Base Model: Z-Image Base
Trigger Word: AvaviLora
Recommended Strength: [e.g., 0.7 - 1.0]

📝 How to Prompt

This LoRA was trained using natural language sentences and long, descriptive paragraphs. It responds best when you describe the scene like a story rather than just using a list of tags.

Prompting Strategy:

Start with the trigger: Always begin with AvaviLora.
Describe the subject: Define gender, age, and action (e.g., "The image shows a female Na'vi aiming a bow...").
Add details: Mention physical traits (e.g., "blue skin, bioluminescent dots, yellow eyes, braided hair with beads, war paint").
Set the scene: Describe the lighting and background (e.g., "cinematic lighting, blurred jungle background, intense atmosphere").

Key Capabilities:

Physiology: blue skin, yellow eyes, fangs, pointed ears, bioluminescent dots.
Accessories: war paint (green/yellow stripes), feathers, wooden beads, leather headpiece.
Expressions: fierce, cautious, tender, screaming, focused.
Environments: lush forest, bioluminescent night, underwater, flying on ikran.

🎨 Example Prompt

AvaviLora, The image shows a female Na'vi with a cautious and intense expression as she appears to be sneaking. Her face is painted with vibrant war paint patterns in shades of green and yellow, contrasting against her blue skin. Her eyes are wide open, gazing intently. The lighting creates an atmosphere of suspense, illuminating parts of her face with warm tones against a blurred jungle background.

⚙️ Recommended Settings

Sampler: res multistep
Steps: ~35
CFG Scale: 3-4
Resolution: 1920x1088 (best case)

Avatar Navi Z-Image Base - v1