MNIST Neural Network is a python application that trains a neural network on the MNIST dataset to recognize handwritten digits WITHOUT using any machine learning libraries (only NumPy). The neural network contains only one fully connected layer with 1024 neurons with ReLU activation function and an output layer with softmax activation function, and the model is trained using the cross-entropy loss function. This is one of two parts of the interactive web application, which you can test out: Live Demo.
These instructions will give you a copy of the neural network up and running on your local machine for development and testing purposes.
To run this application locally on your computer, you'll need Git and Python installed on your computer.
Then run the following command in the command line and go to the desired directory to store this project:
Clone this repository:
git clone https://github.com/JonathanCen/mnist-neural-network.git
Start running the neural network:
python3 main
- NumPy - Used to for all the matrix computations/manipulated
All issues and feature requests are welcome. Feel free to check the issues page if you want to contribute.
Copyright © 2019 Jonathan Cen.
This project is MIT licensed.