Naja-Portrait-V1

Model Details

Model Name: Naja-Portrait-V1
Version: 1.0
Date Created: 2025-09-30
Authors: Natthawut Suthongsa
Architecture: Stable Diffusion 1.5-based LoRA
Parameters: ~1.2M (LoRA)
Input Type: Text prompt
Output Type: Image (512x512)

Intended Use

Positive: Generate realistic portraits of a fictional character "naja_person", artistic, cinematic, vintage-style images
Negative: Not for real identifiable people, offensive content, or modern clothing

Training Data

Mix of public domain portraits and artistic photography (~50k images), no private data

Limitations

Distorted anatomy if prompt unclear
Struggles with modern clothing or multiple subjects

Ethical Considerations

Avoid generating images of real people
Use negative prompts to prevent artifacts

Example Prompt

Positive: "a cinematic, ultra realistic portrait of naja_person, early 20th century, vintage fashion, soft lighting, foggy London streets, tweed coat, flat cap, dramatic shadows"
Negative: "blurry, low quality, cartoon, modern clothes, distorted face, extra limbs"

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support