lovis93/vintage-comic-style-zimage-lora

Vintage Comic Style LoRA for Z-Image Turbo

A LoRA (Low-Rank Adaptation) fine-tuned on Z-Image Turbo to generate images in a distinctive vintage comic / retro illustration style.

🎨 Style Characteristics

This LoRA captures the essence of vintage comic book and screen print illustration with:

• Bold black outlines and strong graphic shapes
• Halftone/stipple textures reminiscent of old print techniques
• Vibrant retro color palettes with limited but striking colors
• Flat color fills with decorative patterns
• 1960s-70s illustration aesthetic

🚀 Usage

Trigger Word

Use St652yl3 at the beginning of your prompt to activate the style.

Prompt Format

St652yl3, [your scene description]

Example Prompts

St652yl3, a tropical beach with palm trees and crystal clear water
St652yl3, a viking warrior standing proud with an axe, mountains in the background
St652yl3, a retro robot in a futuristic city
St652yl3, centaur archer in mystical forest clearing, large trees dominate the left side

🔗 Run on fal.ai

This LoRA works best with Z-Image Turbo on fal.ai:

Text-to-Image: https://fal.ai/models/fal-ai/z-image/turbo/lora

Image-to-Image: https://fal.ai/models/fal-ai/z-image/turbo/image-to-image/lora

API Usage with fal.ai

import fal_client

result = fal_client.subscribe(
    "fal-ai/z-image/turbo/lora",
    arguments={
        "prompt": "St652yl3, a mystical forest with ancient trees",
        "loras": [
            {
                "path": "https://huggingface.co/lovis93/vintage-comic-style-zimage-lora/resolve/main/vintage_comic_style_lora.safetensors",
                "scale": 1.0
            }
        ]
    }
)

📸 Example Outputs

📊 Training Details

• Base Model: Z-Image Turbo (6B parameters) by Tongyi-MAI
• Training Platform: fal.ai Z-Image Trainer
• Format: SafeTensors
• Tested on: Z-Image Turbo ✅ (best), Flux 2, Qwen Image

💡 Tips

• The LoRA works best with Z-Image Turbo compared to other models (tested on Flux 2, Qwen Image)
• Z-Image is super fast - ideal for rapid iteration
• Works well for landscapes, characters, objects, and scenes
• The style naturally adds the vintage comic aesthetic to any subject
• Recommended scale: 1.0

📄 License

This LoRA is released under the Apache 2.0 license for commercial and non-commercial use.

🙏 Credits

• Base model: Z-Image Turbo by Tongyi-MAI
• Training infrastructure: fal.ai
• Style inspiration: Vintage comic books and screen print illustrations