總之就是用 c 刻的類神經網路，c 比 c++ 刻起來麻煩太多了 = =
It is a lightweight neural network framework implemented in C. It provides functionalities to create, train, and use neural networks for various machine learning tasks.
- Support for multiple activation functions (ReLU, Sigmoid, Tanh, Linear, Leaky ReLU).
- Different loss functions available (MSE, Cross Entropy, Binary Cross Entropy, MAE).
- Easily customizable network architecture with adjustable layer sizes and configurations.
- Includes functions for training with backpropagation and predicting outputs.
-
Clone the repository:
git clone https://github.com/afan0918/Neural-Network.git cd Neural-Network -
Compile the framework:
mkdir build cd build cmake .. make
#include "neural_network.h"
// xor test
int main() {
int layers[] = {2, 6, 1};
NeuralNetwork *network = create_network(3, layers, SIGMOID, CROSS_ENTROPY);
initialize_weights(network, -1.0, 1.0);
// XOR training data
double inputs[4][2] = {
{0.0, 0.0},
{0.0, 1.0},
{1.0, 0.0},
{1.0, 1.0}
};
double outputs[4][1] = {
{0.0},
{1.0},
{1.0},
{0.0}
};
int epochs = 10000;
double learning_rate = 0.1;
// Training loop
for (int epoch = 0; epoch < epochs; epoch++) {
double total_loss = 0.0;
for (int i = 0; i < 4; i++) {
forward_propagation(network, inputs[i]);
total_loss += compute_error(network->layers[network->num_layers - 1].neurons[0].output, outputs[i][0]);
backward_propagation(network, outputs[i], learning_rate);
}
if (epoch % 1000 == 0) {
printf("Epoch %d, Loss: %f\n", epoch, total_loss / 4);
for (int i = 0; i < 4; i++) {
forward_propagation(network, inputs[i]);
printf("Input: [%f, %f], Predicted Output: %f, Expected Output: %f\n",
inputs[i][0], inputs[i][1], network->layers[network->num_layers - 1].neurons[0].output, outputs[i][0]);
}
}
}
return 0;
}