Package ai.djl.training.loss

Class MaskedSoftmaxCrossEntropyLoss

java.lang.Object

ai.djl.training.evaluator.Evaluator

ai.djl.training.loss.Loss

ai.djl.training.loss.MaskedSoftmaxCrossEntropyLoss

public class MaskedSoftmaxCrossEntropyLoss
extends Loss

MaskedSoftmaxCrossEntropyLoss is an implementation of Loss that only considers a specific number of values for the loss computations, and masks the rest according to the given sequence.

Field Summary

Fields inherited from class ai.djl.training.evaluator.Evaluator

totalInstances

Constructor Summary

Constructors

Constructor	Description
MaskedSoftmaxCrossEntropyLoss()	Creates a new instance of SoftmaxCrossEntropyLoss with default parameters.
MaskedSoftmaxCrossEntropyLoss(java.lang.String name)	Creates a new instance of SoftmaxCrossEntropyLoss with default parameters.
MaskedSoftmaxCrossEntropyLoss(java.lang.String name, float weight, int classAxis, boolean sparseLabel, boolean fromLogit)	Creates a new instance of MaskedSoftmaxCrossEntropyLoss with the given parameters.

Method Summary

Modifier and Type	Method	Description
NDArray	evaluate(NDList labels, NDList predictions)	Calculates the evaluation between the labels and the predictions.

Methods inherited from class ai.djl.training.loss.Loss

addAccumulator, elasticNetWeightedDecay, elasticNetWeightedDecay, elasticNetWeightedDecay, elasticNetWeightedDecay, getAccumulator, hingeLoss, hingeLoss, hingeLoss, l1Loss, l1Loss, l1Loss, l1WeightedDecay, l1WeightedDecay, l1WeightedDecay, l2Loss, l2Loss, l2Loss, l2WeightedDecay, l2WeightedDecay, l2WeightedDecay, maskedSoftmaxCrossEntropyLoss, maskedSoftmaxCrossEntropyLoss, maskedSoftmaxCrossEntropyLoss, quantileL1Loss, quantileL1Loss, resetAccumulator, sigmoidBinaryCrossEntropyLoss, sigmoidBinaryCrossEntropyLoss, sigmoidBinaryCrossEntropyLoss, softmaxCrossEntropyLoss, softmaxCrossEntropyLoss, softmaxCrossEntropyLoss, updateAccumulator

Methods inherited from class ai.djl.training.evaluator.Evaluator

checkLabelShapes, checkLabelShapes, getName

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MaskedSoftmaxCrossEntropyLoss

public MaskedSoftmaxCrossEntropyLoss()

Creates a new instance of SoftmaxCrossEntropyLoss with default parameters.

MaskedSoftmaxCrossEntropyLoss

public MaskedSoftmaxCrossEntropyLoss(java.lang.String name)

Creates a new instance of SoftmaxCrossEntropyLoss with default parameters.

Parameters:

name - the name of the loss

MaskedSoftmaxCrossEntropyLoss

public MaskedSoftmaxCrossEntropyLoss(java.lang.String name,
                                     float weight,
                                     int classAxis,
                                     boolean sparseLabel,
                                     boolean fromLogit)

Creates a new instance of MaskedSoftmaxCrossEntropyLoss with the given parameters.

Parameters:

name - the name of the loss

weight - the weight to apply on the loss value, default 1

classAxis - the axis that represents the class probabilities, default -1

sparseLabel - whether labels are 1-D integer array of [batch_size] (false) or 2-D probabilities of [batch_size, n-class] (true), default true

fromLogit - if true, the inputs are assumed to be the numbers before being applied with softmax. Then logSoftmax will be applied to input, default false

Method Detail

evaluate

public NDArray evaluate(NDList labels,
                        NDList predictions)

Calculates the evaluation between the labels and the predictions. The label parameter is an NDList that contains the label and the mask sequence in that order.

Specified by:

evaluate in class Evaluator

Parameters:

labels - the NDList that contains correct values and the mask sequence

predictions - the predicted values

Returns:

the evaluation result