README.md

metadata

title: Vision Web App
emoji: 📷
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false

Vision Web App - Object Detection Demo

A multi-model object detection and image classification demo using YOLOv8, DETR, and ViT models. This project is designed to showcase different computer vision models for a hiring demonstration.

Project Architecture

This project follows a phased development approach:

Phase 0: PoC with Gradio (Original)

• Simple Gradio interface with multiple object detection models
• Uses Hugging Face's free tier for model hosting
• Easy to deploy to Hugging Face Spaces

Phase 1: Service Separation (Implemented)

• Backend: Flask API with model inference endpoints
• REST API endpoints for model inference
• JSON responses with detection results and performance metrics

Phase 2: UI Upgrade (Implemented)

• Modern React frontend with Material-UI components
• Improved user experience with responsive design
• Separate frontend and backend architecture

Phase 3: CI/CD & Testing (Planned)

• GitHub Actions for automated testing and deployment
• Comprehensive test suite with pytest and ESLint
• Automatic rebuilds on Hugging Face Spaces

How to Run

Option 1: Original Gradio App

1. Install dependencies:

   pip install -r requirements.txt
   

2. Run the Gradio app:

   python app.py
   

3. Open your browser and go to the URL shown in the terminal (typically http://127.0.0.1:7860)

Option 2: React Frontend with Flask Backend

1. Install backend dependencies:

   pip install -r requirements.txt
   

2. Start the Flask backend server:

   python api.py
   

3. In a separate terminal, navigate to the frontend directory:

   cd frontend
   

4. Install frontend dependencies:

   npm install
   

5. Start the React development server:

   npm start
   

6. Open your browser and go to http://localhost:3000

Models Used

• YOLOv8: Fast and accurate object detection
• DETR: DEtection TRansformer for object detection
• ViT: Vision Transformer for image classification

API Endpoints

The Flask backend provides the following API endpoints:

• GET /api/status - Check the status of the API and available models
• POST /api/detect/yolo - Detect objects using YOLOv8
• POST /api/detect/detr - Detect objects using DETR
• POST /api/classify/vit - Classify images using ViT

All POST endpoints accept form data with an 'image' field containing the image file.

Deployment

Gradio App

The Gradio app is designed to be easily deployed to Hugging Face Spaces:

1. Create a new Space on Hugging Face
2. Select Gradio as the SDK
3. Push this repository to the Space's git repository
4. The app will automatically deploy

React + Flask App

For the React + Flask version, you'll need to:

1. Build the React frontend:

   cd frontend
   npm run build
   

2. Serve the static files from a web server or cloud hosting service

3. Deploy the Flask backend to a server that supports Python