org.deeplearning4j.spark.earlystopping

Class SparkDataSetLossCalculator

java.lang.Object

org.deeplearning4j.spark.earlystopping.SparkDataSetLossCalculator

All Implemented Interfaces:

java.io.Serializable, ScoreCalculator<MultiLayerNetwork>

public class SparkDataSetLossCalculator
extends java.lang.Object
implements ScoreCalculator<MultiLayerNetwork>

Score calculator to calculate the total loss for the MultiLayerNetwork on that data set (data set as a JavaRDD), using Spark. Typically used to calculate the loss on a test set.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
SparkDataSetLossCalculator(org.apache.spark.api.java.JavaRDD<org.nd4j.linalg.dataset.DataSet> data, boolean average, org.apache.spark.SparkContext sc) Calculate the score (loss function value) on a given data set (usually a test set)

Method Summary

Modifier and Type	Method and Description
double	calculateScore(MultiLayerNetwork network) Calculate the score for the given MultiLayerNetwork

Methods inherited from class java.lang.Object

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

Constructor Detail

SparkDataSetLossCalculator

public SparkDataSetLossCalculator(org.apache.spark.api.java.JavaRDD<org.nd4j.linalg.dataset.DataSet> data,
                                  boolean average,
                                  org.apache.spark.SparkContext sc)

Calculate the score (loss function value) on a given data set (usually a test set)

Parameters:

data - Data set to calculate the score for

average - Whether to return the average (sum of loss / N) or just (sum of loss)

Method Detail

calculateScore

public double calculateScore(MultiLayerNetwork network)

Description copied from interface: ScoreCalculator

Calculate the score for the given MultiLayerNetwork

Specified by:

calculateScore in interface ScoreCalculator<MultiLayerNetwork>