DigitVision is a flask web application built on PyTorch that allows users to interactively train and visualize a neural network for recognizing handwritten digits. The app provides an intuitive interface for drawing digits, training the model, and visualizing training metrics such as loss and accuracy.

• Interactive Digit Drawing: Draw digits on a grid to test the neural network's predictions.
• Train Induvidual Digits By clicking on the output neuron you can train your own digits.
• Real-Time Training: Train the neural network on a dataset of handwritten digits and visualize training progress in real-time.
• Metrics Visualization: View training and validation loss and accuracy in a popup graph.

Description: The main interface where users can draw digits and interact with the neural network.

Description: Popup window displaying training and validation metrics.

Follow these instructions to set up and run the application using Docker.

Ensure you have the following installed:

• Docker
• Docker Compose

1. Clone the repository:

   git clone https://github.com/dkruyt/digitvision.git
   cd digitvision

2. Build and run the application using Docker Compose:

   docker-compose up --build

   This will build the Docker image and start the application. The app will be available at http://localhost:8000.

1. Open your browser and navigate to:

   http://localhost:8000
   

2. Interact with the App:

   • Use the input grid to draw a digit.
   • Click the "Predict" button to see the neural network's prediction.
   • Load training data and start training the model using the provided buttons.
   • View training and validation metrics by clicking the "Graphs" button.
   • Save and load trained models using the respective buttons.

digitvision/
│
├── Dockerfile                # Dockerfile for building the Docker image
├── docker-compose.yaml       # Docker Compose configuration
├── entrypoint.sh             # Entrypoint script for Docker container
├── requirements.txt          # Python dependencies
├── server.py                 # Flask server and neural network training logic
├── static/
│   └── index.js              # JavaScript for client-side functionality
└── templates/
    └── index.html            # Main HTML file

The docker-compose.yaml file is configured to build and run the application with the following default environment variables:

• HIDDEN_NEURONS: Number of neurons in the hidden layer (default: 8)
• LIMIT_PER_DIGIT: Number of digits per class for training (default: 16)
• NUM_CLASSES: Number of output neurons and number of digit classes (default: 10)

These values can be adjusted by editing the docker-compose.yaml file.

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a new branch (git checkout -b feature-branch).
3. Make your changes.
4. Commit your changes (git commit -am 'Add new feature').
5. Push to the branch (git push origin feature-branch).
6. Create a new Pull Request.

This project is licensed under the GPLv3 License - see the LICENSE file for details.

• The Optical Recognition of Handwritten Digits dataset used in this app is provided by the UCI Machine Learning Repository.
• Thanks to the open-source community for providing the tools and libraries that made this project possible.
• Sparky, for his BackPropagation app, which inspired this application.

Happy training with DigitVision!