CNeuralNet is a neural network library implemented from scratch in C for image classification tasks using the MNIST dataset. The library provides tools to build, train, and evaluate neural networks, and it is optimized for efficient performance. The neural network architecture includes fully connected layers with customizable activation functions, and it uses the backpropagation algorithm for training.
-Fully connected neural network with customizable architecture.
-Activation functions: Sigmoid and Softmax.
-Backpropagation algorithm for training.
-Efficient and optimized C implementation.
This neural network trained on the popular MNIST dataset. By leveraging C's efficiency and optimization capabilities, the network achieves an accuracy of 89% in classifying handwritten digits from 0 to 9.