Method Infer

src/RaceNet/RaceNet.cpp:22–35

RaceNet::RaceNet() : net(layerParams, learningRate) {}

vector<double> RaceNet::Infer(vector<double> raycastInputs) {
        // as the sigmoid function never reaches 0.0 nor 1.0
        // it can be a good idea to consider values greater than 0.9 as 1.0 and values smaller than 0.1 as 0.0
        // hence the step function.
        Matrix<double> resultMatrix = net.computeOutput(raycastInputs);//.applyFunction(stepFunction);
        std::vector<double> resultVector;

        resultVector.emplace_back(resultMatrix.get(0, 0));
        resultVector.emplace_back(resultMatrix.get(0, 1));
        resultVector.emplace_back(resultMatrix.get(0, 2));
        //resultVector.emplace_back(resultMatrix.get(0, 3));

        return resultVector;
}