cpp-neural-network

Configurable Feed-Forward Backpropagation Neural Network written in C++ from Scratch.

This implementation supports arbitrary depth feed-forward backpropagation neural networks.

Configurable Parameters :

• Number of Layers
• Number of neurons in each Layer
• Activation Function for each Layer (ReLU, Sigmoid, TanH, Softmax)
• Network Architecture (Fully Connected / Partially Connected)
• L2 Regularization Strength to Avoid Overfitting

Usage :

• Clone the repository
• Compile and run neural_network.cpp
• Enter asked inputs
• Training starts inputted iterations
• Check predicted outputs and accuracy in prediction.txt