This repository contains a project developed for Hack Club. It implements a Convolutional Neural Network (CNN) to recognize handwritten digits from the MNIST dataset. For an introduction to neural networks, see this guide.
- Python >= 3.8
Setting Up the Environment
- Windows:
./start_Windows.bat
- Linux/macOS:
./start_Linux.sh
These scripts will install required dependencies, and build a virtual environment for you if you don't have one. They will also start the GUI automatically
CLI:
-
Navigate to the
bin
directory:cd bin
-
Choose your method:
- GUI:
python MNIST_gui.py [--help]
(usepython3
on Linux/macOS)
This will launch the GUI for testing the neural network, it will open the browser automatically.
You can find examples of images in the data/Examples folder.
- Model:
python MNIST.py [--help]
(usepython3
on Linux/macOS)
This will run the model via command line interface.
The
--help
flag displays available command-line arguments. - GUI:
GUI:
Double click on the start files for your OS, and it should open the GUI app automatically.
The GUI allows for an interactive experience to test the neural network. Follow the on-screen instructions and see the model's predictions in real-time.
The command-line interface allows for training and testing the model through specified arguments. Use the --help flag to explore available options.
Neetre