Files
GitHub Action d77f071287 🎉 Initial release: ComfyUI-OmniSVG v1.0.0
Features:
- 🔧 OmniSVG Model Loader - Load and manage OmniSVG models
- 📝 Text to SVG - Generate SVG graphics from text descriptions
- 🖼 Image to SVG - Convert images to SVG format
- 🔄 SVG to Image - Convert SVG strings to ComfyUI IMAGE tensors
- 💾 SVG Saver - Save SVG files to disk

- 100% working implementation with OmniSVG-3B model
- Complete ComfyUI workflow integration
- Professional documentation and examples
- Automated installation and testing
- Memory management for 17GB VRAM models

- All 5 functional nodes
- Example workflows and documentation
- Automated installer and test suite
- GitHub Actions CI/CD pipeline
- Comprehensive troubleshooting guides

Ready for production use! �
2025-08-02 21:42:23 +00:00
..
2025-08-02 21:42:23 +00:00

ComfyUI-OmniSVG Examples

This directory contains example workflows and usage patterns for ComfyUI-OmniSVG.

📁 Files

basic_workflow.json

Complete workflow demonstrating all OmniSVG nodes:

  • Model loading
  • Text-to-SVG generation
  • Image-to-SVG conversion
  • SVG processing and saving
  • Preview integration

🚀 How to Use Examples

  1. Load Workflow in ComfyUI

    • Open ComfyUI
    • Click "Load" button
    • Select the workflow JSON file
    • All nodes will be automatically arranged
  2. Configure Parameters

    • Adjust text prompts for different SVG styles
    • Modify generation parameters (temperature, top_p, etc.)
    • Set output directories and filenames
  3. Run Generation

    • Click "Queue Prompt" to start generation
    • Monitor progress in the console
    • View results in preview nodes

💡 Example Prompts

Text-to-SVG Prompts

"A minimalist icon of a coffee cup with steam"
"Simple geometric pattern with triangles and circles"
"Logo design for a tech company with clean lines"
"Weather icon showing sun and clouds"
"Navigation arrow pointing right"

Image-to-SVG Tips

  • Use simple, high-contrast images for best results
  • Logos and icons work particularly well
  • Resize images to 200-400px for optimal processing
  • Black and white images often produce cleaner SVGs

🎯 Workflow Patterns

Pattern 1: Text-to-SVG Pipeline

Model Loader → Text to SVG → SVG to Image → Preview
                     ↓
                 SVG Saver

Pattern 2: Image-to-SVG Pipeline

Load Image → Model Loader → Image to SVG → SVG to Image → Preview
                                  ↓
                              SVG Saver

Pattern 3: Batch Processing

Model Loader (shared)
    ↓
Multiple Text/Image to SVG nodes
    ↓
Multiple SVG processors

🔧 Customization

Generation Parameters

  • Temperature: Controls randomness (0.1-2.0)
    • Lower = more consistent
    • Higher = more creative
  • Top-p: Nucleus sampling (0.1-1.0)
  • Top-k: Token selection limit (1-100)
  • Repetition Penalty: Prevents repetition (1.0-2.0)

Output Settings

  • SVG Dimensions: Modify in SVG to Image node
  • File Naming: Customize in SVG Saver node
  • Output Directory: Set custom paths

📊 Performance Tips

  1. GPU Memory: Ensure 17GB+ VRAM available
  2. Batch Size: Process one at a time for stability
  3. Model Loading: Reuse loaded models across generations
  4. Memory Cleanup: Restart ComfyUI if memory issues occur

🐛 Troubleshooting Examples

Common Issues

  1. "Model not found": Check model path in loader
  2. "CUDA out of memory": Free GPU memory or use CPU
  3. "Generation failed": Try different parameters
  4. "SVG invalid": Check input image quality

Debug Workflow

  1. Load basic_workflow.json
  2. Use simple text prompt: "red circle"
  3. Check each node output
  4. Verify model loading first

🎨 Creative Ideas

Icon Sets

Create consistent icon sets by using similar prompts:

  • "Minimalist home icon"
  • "Minimalist settings icon"
  • "Minimalist user icon"

Logo Variations

Generate logo variations:

  • "Tech company logo with geometric shapes"
  • "Same logo but with rounded corners"
  • "Same logo but monochrome"

Pattern Generation

Create repeatable patterns:

  • "Seamless geometric pattern"
  • "Art deco border design"
  • "Simple repeating motif"

📝 Notes

  • All examples are tested and working
  • Modify parameters to suit your needs
  • Share your own workflows with the community
  • Report issues or improvements on GitHub