ai.djl.training.metrics

Class Accuracy

java.lang.Object

ai.djl.training.metrics.TrainingMetric

ai.djl.training.metrics.Accuracy

All Implemented Interfaces:

java.lang.Cloneable

Direct Known Subclasses:

SingleShotDetectionAccuracy, TopKAccuracy

public class Accuracy
extends TrainingMetric

Accuracy is a TrainingMetric that computes the accuracy score.

The accuracy score is defined as \(accuracy(y, \hat{y}) = \frac{1}{n}\sum_{i=0}^{n-1}1(\hat{y_i} == y_i)\)

Field Summary

Fields

Modifier and Type	Field and Description
protected int	axis
protected int	index

Constructor Summary

Constructors

Constructor and Description
Accuracy() Creates an accuracy metric that computes accuracy across axis 1 along the 0th index.
Accuracy(java.lang.String name, int index) Creates an accuracy metric that computes accuracy across axis 1 along given index.
Accuracy(java.lang.String name, int index, int axis) Creates an accuracy metric.

Method Summary

Modifier and Type	Method and Description
void	addCorrectInstances(long numInstances) Add a number to the correct instances.
void	addTotalInstances(long totalInstances) Add a number to the total instances.
float	getValue() Calculates metric values.
void	reset() Resets metric values.
void	update(NDArray labels, NDArray predictions) Computes and updates the accuracy based on the labels and predictions.
void	update(NDList labels, NDList predictions) Computes and updates the training metrics based on the labels and predictions.

Methods inherited from class ai.djl.training.metrics.TrainingMetric

checkLabelShapes, checkLabelShapes, duplicate, getName

Methods inherited from class java.lang.Object

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

Field Detail

axis

protected int axis

index

protected int index

Constructor Detail

Accuracy

public Accuracy(java.lang.String name,
                int index,
                int axis)

Creates an accuracy metric.

Parameters:

name - the name of the metric, default is "Accuracy"

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

axis - the axis that represent classes in prediction, default 1

Accuracy

public Accuracy()

Creates an accuracy metric that computes accuracy across axis 1 along the 0th index.

Accuracy

public Accuracy(java.lang.String name,
                int index)

Creates an accuracy metric that computes accuracy across axis 1 along given index.

Parameters:

name - the name of the metric, default is "Accuracy"

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

Method Detail

reset

public void reset()

Resets metric values.

Specified by:

reset in class TrainingMetric

update

public void update(NDArray labels,
                   NDArray predictions)

Computes and updates the accuracy based on the labels and predictions.

Parameters:

labels - a NDList of labels

predictions - a NDList of predictions

update

public void update(NDList labels,
                   NDList predictions)

Computes and updates the training metrics based on the labels and predictions.

Specified by:

update in class TrainingMetric

Parameters:

labels - a NDList of labels

predictions - a NDList of predictions

getValue

public float getValue()

Calculates metric values.

Specified by:

getValue in class TrainingMetric

Returns:

a Pair of metric name and value

addCorrectInstances

public void addCorrectInstances(long numInstances)

Add a number to the correct instances.

Parameters:

numInstances - the number to increment by

addTotalInstances

public void addTotalInstances(long totalInstances)

Add a number to the total instances.

Parameters:

totalInstances - the number to increment by