AYA Vision Detection Demo

A Flask-based web application that demonstrates the capabilities of Cohere's c4ai-aya-vision-32b model for object detection in images.

Features

• Configurable Detection Subject: Set what you want to detect in images (e.g., flares, buildings, vehicles)
• Image Upload: Upload multiple images for batch processing
• Image Analysis: Process images using Cohere's c4ai-aya-vision-32b model
• Results Display: View detection results in a grid layout with thumbnails
• Full-Size Image Viewing: Click on thumbnails to view full-size images
• Image Deletion: Delete images from both grid and full-size views
• Filtering and Sorting: Filter and sort images by detection status
• Responsive Design: Works on desktop and mobile devices

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/aya-vision-demo.git
   cd aya-vision-demo
   

2. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install the required packages:

   pip install -r requirements.txt
   

4. Set up your Cohere API key:

   export COHERE_API_KEY=your_api_key_here
   

   On Windows:

   set COHERE_API_KEY=your_api_key_here
   

Usage

1. Start the application:

   python run.py
   

2. Open your web browser and navigate to:

   http://localhost:5001
   

3. Configure the detection subject and prompt in the Settings page

4. Upload images for analysis

5. View and interact with the results

Project Structure

aya-vision-demo/
├── app/
│   ├── __init__.py
│   ├── config.py
│   ├── forms.py
│   ├── routes.py
│   ├── utils.py
│   ├── static/
│   │   ├── css/
│   │   │   └── style.css
│   │   ├── js/
│   │   │   └── main.js
│   │   └── uploads/
│   └── templates/
│       ├── base.html
│       ├── index.html
│       ├── results.html
│       └── settings.html
├── requirements/
│   └── prd.md
├── requirements.txt
└── run.py

Technologies Used

• Flask: Web framework
• Cohere API: For image analysis using the c4ai-aya-vision-32b model
• Bootstrap 5: For responsive design
• Font Awesome: For icons
• JavaScript: For interactive features

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgements

• Cohere for providing the c4ai-aya-vision-32b model
• Flask community for the excellent web framework