We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Abubakar Abid - Keynote: Building Machine Learning Apps in Python with Gradio | PyData Global 2023
Learn how to build web interfaces for ML models using Gradio with PyData keynote speaker Abubakar Abid. Covers Interface & Blocks APIs, demos, and new 4.0 features.
-
Gradio is a Python library that makes it easy to create web interfaces for machine learning models with just a few lines of code
-
Key components needed for a Gradio interface:
- A Python function
- Input component(s) that map to function parameters
- Output component(s) that display function results
-
Gradio offers two main APIs:
- Interface API: High-level, quick demos with minimal code
- Blocks API: Low-level, full control over layout and processing
-
Every Gradio app automatically gets:
- A web UI accessible via browser
- API endpoint for programmatic access
- Built-in queuing system for handling concurrent users
- Share links for easy demo sharing
-
Common use cases:
- Image classification demos
- Text generation interfaces
- Audio/speech processing
- Real-time dashboards
- Interactive machine learning model demos
-
Key benefits:
- Makes ML models accessible to non-technical users
- Helps identify biases and gaps in model performance
- Enables quick feedback from users
- Refocuses research on practical applications
-
Hosting options:
- Local deployment
- Hugging Face Spaces (free tier available)
- Custom servers
-
New in Gradio 4.0:
- Improved performance via Svelte frontend
- Custom component creation
- Enhanced theme customization
- Better streaming support
- Optional queuing system