ksteinbraker/coloring-book-ai-generator

🎨 Photo to Coloring Book Generator

This Hugging Face Space transforms your photos into beautiful coloring book pages using AI-powered edge detection and line art generation.

Features

• Upload any photo and get a coloring page version
• Choose styles: Clean Lines, Anime/Manga, Soft Edges
• Complexity levels: Simple (Kids), Medium, Detailed (Adults)
• Advanced controls for guidance and steps
• Fallback to basic edge detection if AI models fail
• Updated to Gradio 5.36.2 for the latest security fixes, features, and to resolve SDK version warnings.

How It Works

1. Preprocesses the image using ControlNet's Canny detector or basic OpenCV
2. Uses Stable Diffusion 1.5 with ControlNet Lineart to generate clean outlines
3. Falls back gracefully on errors

Requirements

See requirements_stable.txt for dependencies.

Deployment Instructions

1. Create a new HF Space (don't reuse old ones to avoid cache issues).
2. Upload these exact files: app_stable.py, requirements_stable.txt, and this README.md.
3. If uploading via web: Drag and drop into the Files tab.
4. For Git: Clone your Space repo, add files, commit, and push.
5. Monitor build logs for any install errors.
6. Test with a simple photo; check Space logs for runtime issues.

Troubleshooting

• Cached Download Errors: Ensure huggingface-hub and diffusers versions match (pinned in requirements).
• OOM Issues: The a10g-small GPU should suffice; if not, try cpu-upgrade as fallback.
• Slow Loading: Models load on first use; initial run may take longer.
• If issues persist, check HF status or share your Space URL/logs.
• If you encounter import errors with diffusers, ensure you're using diffusers==0.24.0 and torch==2.0.1+cu117.

Built with ❤️ using Grok's guidance.