Package ai.djl.training.evaluator

package ai.djl.training.evaluator

Contains classes for evaluating the effectiveness of models.

It contains a main interface Evaluator and various evaluators that extend it. More evaluators are located within ai.djl.training.loss which have the additional property that those evaluators are suited for training.

Related Packages

Package	Description
ai.djl.training	Contains classes and implementations to train a neural network.

Classes

Class	Description
AbstractAccuracy	Accuracy is an Evaluator that computes the accuracy score.
Accuracy	Accuracy is the AbstractAccuracy with multiple classes.
BinaryAccuracy	BinaryAccuracy is the AbstractAccuracy with two classes.
BoundingBoxError	BoundingBoxError is an Evaluator that computes the error in the prediction of bounding boxes in SingleShotDetection model.
Coverage	Coverage for a Regression problem: it measures the percent of predictions greater than the actual target, to determine whether the predictor is over-forecasting or under-forecasting.
Evaluator	Base class for all Evaluators that can be used to evaluate the performance of a model.
IndexEvaluator	A wrapper for an Evaluator that evaluates on only a particular NDArray in the predictions and/or labels NDLists.
SingleShotDetectionAccuracy	SingleShotDetectionAccuracy is an implementation of AbstractAccuracy.
TopKAccuracy	TopKAccuracy is an Evaluator that computes the accuracy of the top k predictions.