Basic Neural Network Training in C

This project demonstrates a basic implementation of a neural network in C programming language. The neural network is trained using the MNIST dataset to recognize handwritten digits.

Requirements

• C compiler (e.g., GCC)
• Make (for compilation)
• MNIST dataset
• Linux environment (recommended); if using Windows, WSL (Windows Subsystem for Linux) is recommended due to potential issues with floating-point accumulation.

Compilation

Compile the project using make:

make

Get the training dataset

kaggle datasets download -d oddrationale/mnist-in-csv
unzip mnist-in-csv.zip -d data