【deep learning学习笔记】注释yusugomori的DA代码 --- dA.cpp -- 模型测试

来源：互联网发布：win10默认网卡mac地址编辑：程序博客网时间：2024/05/22 17:41

测试代码。能看到，训练的时候是单个样本、单个样本的训练的，在NN中是属于“stochastic gradient descent”，否则，一批样本在一起的，就是“standard gradient descent”。

void test_dA() {srand(0);  double learning_rate = 0.1;double corruption_level = 0.3;int training_epochs = 100;int train_N = 10;int test_N = 2;int n_visible = 20;int n_hidden = 5;// training dataint train_X[10][20] = {{1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0}};// construct dAdA da(train_N, n_visible, n_hidden, NULL, NULL, NULL);// trainfor(int epoch=0; epoch<training_epochs; epoch++) {// train it sample by samplefor(int i=0; i<train_N; i++) {da.train(train_X[i], learning_rate, corruption_level);}}// test dataint test_X[2][20] =   {{1, 0, 1, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},{0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1, 0}  };double reconstructed_X[2][20];// testfor(int i=0; i<test_N; i++) {da.reconstruct(test_X[i], reconstructed_X[i]);for(int j=0; j<n_visible; j++) {printf("%.5f ", reconstructed_X[i][j]);}cout << endl;} cout << endl;}int main() {test_dA();getchar();return 0;}

程序运行结果：