README.md

metadata

title: TangoBot
emoji: 📸
colorFrom: red
colorTo: yellow
sdk: gradio
python_version: '3.12'
app_file: app.py
pinned: false

TangoBot

Gradio web UI for face detection and age estimation, designed for HuggingFace Spaces deployment. This demo showcases TangoBot's photo analysis capabilities for Tango Charities.

About Tango Charities

Tango Charities is a volunteer-led nonprofit fighting hunger through Feed The City events. Since 2015, they've mobilized 110,000+ volunteers to provide 8+ million meals across 73 cities. TangoBot helps the team count attendees in event photos and identify youth participation.

Features

• Interactive web interface for image upload
• Real-time face detection and age estimation
• Webcam and clipboard support
• Embeds inference directly (no external HTTP calls)

Running Locally

cd apps/huggingface-space
uv run gradio app.py

Opens at http://localhost:7860

Usage

1. Upload an image (or use webcam/clipboard)
2. View the annotated image with face bounding boxes
3. See detection results:
   • Number of faces detected
   • Count of people estimated under 18

Color Coding

Bounding boxes are color-coded:

• Green: Under 18 (minors)
• Blue: 18 and older (adults)

Deploying to HuggingFace Spaces

1. Create a new Space on HuggingFace (Gradio SDK)
2. Copy the package source and app files:

   packages/face-age-inference/src/face_age_inference/ -> face_age_inference/
   apps/huggingface-space/app.py -> app.py
   apps/huggingface-space/requirements.txt -> requirements.txt
   

3. Push to the Space repository

Models are automatically downloaded from HuggingFace Hub on first run (~450MB).

Requirements

See requirements.txt for dependencies. Key packages:

• gradio>=5.0.0 - Web UI framework
• torch>=2.9.1 - PyTorch
• transformers>=4.51.0 - HuggingFace Transformers
• ultralytics>=8.3.230 - YOLO implementation

Configuration

The app uses default settings from face-age-inference. To customize, modify environment variables:

FACE_AGE_DEVICE=cuda:0 uv run gradio app.py