MNIST Digit Recognizer in C

This project implements a simple neural network to recognize handwritten digits from the MNIST dataset using C. It uses ReLU activation in the hidden layer and softmax in the output layer.

Features

• Single hidden layer neural network
• ReLU activation function in the hidden layer
• Softmax activation function in the output layer

Dataset

The project uses the MNIST dataset from Kaggle. Place the following files in the dataset folder:

• train.csv
• test.csv

Compilation

To compile the project, use the following command:

make build

Alternatively you can also run the following command:

gcc main.c neuralNetwork.c utils.c -o mnist_recognizer -lm

This will generate the binary mnist_recognizer.

Running the Program

After compilation, run the program with: ./mnist_recognizer

The program will train on the training data and then generate a sample_submission.csv file with predictions for the test set, which you can submit to Kaggle.

Note

This is a basic implementation and may take a while to run on the full dataset. This can be improved in a thousand ways! Hesitate not to contribute.