Package ai.djl.nn.transformer

Class BertNextSentenceLoss

java.lang.Object

ai.djl.training.evaluator.Evaluator

ai.djl.training.loss.Loss

ai.djl.nn.transformer.BertNextSentenceLoss

public class BertNextSentenceLoss
extends Loss

Calculates the loss for the next sentence prediction task.

Field Summary

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

totalInstances

Constructor Summary

Constructors

Constructor	Description
BertNextSentenceLoss(int labelIdx, int nextSentencePredictionIdx)	Creates a new bert next sentence loss.

Method Summary

Modifier and Type	Method	Description
NDArray	accuracy(NDList labels, NDList predictions)	Calculates the fraction of correct predictions.
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

BertNextSentenceLoss

public BertNextSentenceLoss(int labelIdx,
                            int nextSentencePredictionIdx)

Creates a new bert next sentence loss.

Parameters:

labelIdx - index of the next sentence labels

nextSentencePredictionIdx - index of the next sentence prediction in the bert output

Method Detail

evaluate

public NDArray evaluate(NDList labels,
                        NDList predictions)

Calculates the evaluation between the labels and the predictions.

Specified by:

evaluate in class Evaluator

Parameters:

labels - the correct values

predictions - the predicted values

Returns:

the evaluation result

accuracy

public NDArray accuracy(NDList labels,
                        NDList predictions)

Calculates the fraction of correct predictions.

Parameters:

labels - the labels with the correct predictions

predictions - the bert pretraining model output

Returns:

the fraction of correct predictions.