Stable Diffusion Gradio Space

Features

• Generate images from text prompts
• Customizable seed for reproducible results
• Adjustable width/height
• Public API endpoint at /api/predict
• Optimized for free GPU on HF Spaces

Deploy to Hugging Face Spaces

1. Go to huggingface.co/spaces
2. Click "Create new Space"
3. Fill in:
   • Owner: your username
   • Space name: e.g., sd-turbo-gen
   • License: MIT
   • Hardware: A10G small (free, has GPU)
   • Public: ✓
4. Click "Create Space"
5. In the Files tab, upload these 2 files:
   • app.py
   • requirements.txt
   • (optional) README.md
6. Wait for the space to build and start (10-20 minutes for first model download)
7. Your API endpoint will be: https://your-username-sd-turbo-gen.hf.space/api/predict

Using the API

POST /api/predict

1. prompt (string)
2. seed (number, use -1 for random)
3. width (number, multiple of 64)
4. height (number, multiple of 64)

• data[0]: base64-encoded PNG image
• data[1]: the seed actually used

Client Script

Test locally (optional)

Notes

• First run will download ~2GB model (takes 10-20 min). Subsequent runs are cached.
• The space may go to sleep after inactivity; wake it by visiting the URL.
• Free GPU on HF Spaces has ~16GB RAM, enough for sd-turbo.