ai.djl.training.evaluator

Class TopKAccuracy

java.lang.Object

ai.djl.training.evaluator.Evaluator

ai.djl.training.evaluator.AbstractAccuracy

ai.djl.training.evaluator.TopKAccuracy

public class TopKAccuracy
extends AbstractAccuracy

TopKAccuracy is an Evaluator that computes the accuracy of the top k predictions.

TopKAccuracy differs from AbstractAccuracy in that it considers the prediction to be `True` as long as the ground truth label is in the top K predicated labels. If `top_k = 1`, then TopKAccuracy is identical to Accuracy.

Field Summary

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

axis, correctInstances, index

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

totalInstances

Constructor Summary

Constructors

Constructor and Description
TopKAccuracy(int topK) Creates an instance of TopKAccuracy evaluator that computes topK accuracy across axis 1 along the 0th index.
TopKAccuracy(int index, int topK) Creates an instance of TopKAccuracy evaluator that computes topK accuracy across axis 1 along the given index.
TopKAccuracy(java.lang.String name, int index, int topK) Creates a TopKAccuracy instance.

Method Summary

Modifier and Type	Method and Description
protected ai.djl.util.Pair<java.lang.Long,NDArray>	accuracyHelper(NDList labels, NDList predictions) A helper for classes extending AbstractAccuracy.

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

addAccumulator, evaluate, getAccumulator, resetAccumulator, 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

TopKAccuracy

public TopKAccuracy(java.lang.String name,
                    int index,
                    int topK)

Creates a TopKAccuracy instance.

Parameters:

name - the accuracy name, default "Top_K_Accuracy"

index - the index of the NDArray in labels to compute topK accuracy for

topK - the value of K

TopKAccuracy

public TopKAccuracy(int index,
                    int topK)

Creates an instance of TopKAccuracy evaluator that computes topK accuracy across axis 1 along the given index.

Parameters:

index - the index of the NDArray in labels to compute topK accuracy for

topK - the value of K

TopKAccuracy

public TopKAccuracy(int topK)

Creates an instance of TopKAccuracy evaluator that computes topK accuracy across axis 1 along the 0th index.

Parameters:

topK - the value of K

Method Detail

accuracyHelper

protected ai.djl.util.Pair<java.lang.Long,NDArray> accuracyHelper(NDList labels,
                                                                  NDList predictions)

Description copied from class: AbstractAccuracy

A helper for classes extending AbstractAccuracy.

Specified by:

accuracyHelper in class AbstractAccuracy

Parameters:

labels - the labels to get accuracy for

predictions - the predictions to get accuracy for

Returns:

a pair(number of total values, ndarray int of correct values)