int numHiddenNodes = 20;
bool useBias = true;

nn::Net<float> net;
net.add(new nn::Dense<>(batchSize, numFeatures, numHiddenNodes, useBias));
net.add(new nn::Relu<>());
net.add(new nn::Dense<>(batchSize, numHiddenNodes, numHiddenNodes, useBias));
net.add(new nn::Relu<>());
net.add(new nn::Dense<>(batchSize, numHiddenNodes, numClasses, useBias));
net.add(new nn::Softmax<>());

nn::CrossEntropyLoss<float, 2> lossFunc;
net.registerOptimizer(new nn::Adam<float>(0.01));

int numEpoch = 250;
float loss_t, accuracy_t;
for (unsigned int ii = 0; ii < numEpoch; ++ii)
{
    auto result = net.forward<2, 2>(input);

    float loss = lossFunc.loss(result, labels);
    float accuracy = lossFunc.accuracy(result, labels);
    std::cout << std::setprecision(5);
    std::cout << "Epoch: " << ii << " loss: " << loss << " accuracy: " << accuracy << std::endl;

    auto lossBack = lossFunc.backward(result, labels);
    net.backward(lossBack);
    net.step();
    loss_t = loss;
    accuracy_t = accuracy;
}