Examples with ILossFunction org.nd4j.linalg.lossfunctions.ILossFunction used on opensource projects

Example 16 with ILossFunction

use of org.nd4j.linalg.lossfunctions.ILossFunction in project deeplearning4j by deeplearning4j.

the class CenterLossOutputLayer method computeScoreForExamples.

/**Compute the score for each example individually, after labels and input have been set.
     *
     * @param fullNetworkL1 L1 regularization term for the entire network (or, 0.0 to not include regularization)
     * @param fullNetworkL2 L2 regularization term for the entire network (or, 0.0 to not include regularization)
     * @return A column INDArray of shape [numExamples,1], where entry i is the score of the ith example
     */
@Override
public INDArray computeScoreForExamples(double fullNetworkL1, double fullNetworkL2) {
    if (input == null || labels == null)
        throw new IllegalStateException("Cannot calculate score without input and labels");
    INDArray preOut = preOutput2d(false);
    // calculate the intra-class score component
    INDArray centers = params.get(CenterLossParamInitializer.CENTER_KEY);
    INDArray centersForExamples = labels.mmul(centers);
    INDArray intraClassScoreArray = input.sub(centersForExamples);
    // calculate the inter-class score component
    ILossFunction interClassLoss = layerConf().getLossFn();
    INDArray scoreArray = interClassLoss.computeScoreArray(getLabels2d(), preOut, layerConf().getActivationFn(), maskArray);
    scoreArray.addi(intraClassScoreArray.muli(layerConf().getLambda() / 2));
    double l1l2 = fullNetworkL1 + fullNetworkL2;
    if (l1l2 != 0.0) {
        scoreArray.addi(l1l2);
    }
    return scoreArray;
}

Example 17 with ILossFunction

use of org.nd4j.linalg.lossfunctions.ILossFunction in project deeplearning4j by deeplearning4j.

the class CenterLossOutputLayer method getGradientsAndDelta.

/** Returns tuple: {Gradient,Delta,Output} given preOut */
private Pair<Gradient, INDArray> getGradientsAndDelta(INDArray preOut) {
    ILossFunction lossFunction = layerConf().getLossFn();
    INDArray labels2d = getLabels2d();
    if (labels2d.size(1) != preOut.size(1)) {
        throw new DL4JInvalidInputException("Labels array numColumns (size(1) = " + labels2d.size(1) + ") does not match output layer" + " number of outputs (nOut = " + preOut.size(1) + ")");
    }
    INDArray delta = lossFunction.computeGradient(labels2d, preOut, layerConf().getActivationFn(), maskArray);
    Gradient gradient = new DefaultGradient();
    INDArray weightGradView = gradientViews.get(CenterLossParamInitializer.WEIGHT_KEY);
    INDArray biasGradView = gradientViews.get(CenterLossParamInitializer.BIAS_KEY);
    INDArray centersGradView = gradientViews.get(CenterLossParamInitializer.CENTER_KEY);
    // centers delta
    double alpha = layerConf().getAlpha();
    INDArray centers = params.get(CenterLossParamInitializer.CENTER_KEY);
    INDArray centersForExamples = labels.mmul(centers);
    INDArray diff = centersForExamples.sub(input).muli(alpha);
    INDArray numerator = labels.transpose().mmul(diff);
    INDArray denominator = labels.sum(0).addi(1.0).transpose();
    INDArray deltaC;
    if (layerConf().getGradientCheck()) {
        double lambda = layerConf().getLambda();
        //For gradient checks: need to multiply dLc/dcj by lambda to get dL/dcj
        deltaC = numerator.muli(lambda);
    } else {
        deltaC = numerator.diviColumnVector(denominator);
    }
    centersGradView.assign(deltaC);
    // other standard calculations
    //Equivalent to:  weightGradView.assign(input.transpose().mmul(delta));
    Nd4j.gemm(input, delta, weightGradView, true, false, 1.0, 0.0);
    biasGradView.assign(delta.sum(0));
    gradient.gradientForVariable().put(CenterLossParamInitializer.WEIGHT_KEY, weightGradView);
    gradient.gradientForVariable().put(CenterLossParamInitializer.BIAS_KEY, biasGradView);
    gradient.gradientForVariable().put(CenterLossParamInitializer.CENTER_KEY, centersGradView);
    return new Pair<>(gradient, delta);
}

Example 18 with ILossFunction

use of org.nd4j.linalg.lossfunctions.ILossFunction in project deeplearning4j by deeplearning4j.

the class LossLayer method computeScore.

/** Compute score after labels and input have been set.
     * @param fullNetworkL1 L1 regularization term for the entire network
     * @param fullNetworkL2 L2 regularization term for the entire network
     * @param training whether score should be calculated at train or test time (this affects things like application of
     *                 dropout, etc)
     * @return score (loss function)
     */
@Override
public double computeScore(double fullNetworkL1, double fullNetworkL2, boolean training) {
    if (input == null || labels == null)
        throw new IllegalStateException("Cannot calculate score without input and labels");
    this.fullNetworkL1 = fullNetworkL1;
    this.fullNetworkL2 = fullNetworkL2;
    INDArray preOut = input;
    ILossFunction lossFunction = layerConf().getLossFn();
    //double score = lossFunction.computeScore(getLabels2d(), preOut, layerConf().getActivationFunction(), maskArray, false);
    double score = lossFunction.computeScore(getLabels2d(), preOut, layerConf().getActivationFn(), maskArray, false);
    score += fullNetworkL1 + fullNetworkL2;
    score /= getInputMiniBatchSize();
    this.score = score;
    return score;
}