org.deeplearning4j.arbiter.scoring.util

Class ScoreUtil

java.lang.Object

org.deeplearning4j.arbiter.scoring.util.ScoreUtil

public class ScoreUtil
extends Object

Various utilities for functions used in arbiter.

Author:

Adam Gibson

Constructor Summary

Constructors

Constructor and Description
ScoreUtil()

Method Summary

Modifier and Type	Method and Description
static org.deeplearning4j.eval.Evaluation	getEvaluation(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData) Get the evaluation for the given model and test dataset
static org.deeplearning4j.eval.Evaluation	getEvaluation(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testData) Get the evaluation for the given model and test dataset
static org.deeplearning4j.eval.Evaluation	getEvaluation(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData)
static org.nd4j.linalg.dataset.api.iterator.DataSetIterator	getIterator(Object o) Get a DataSetIterator from the given object whether it's a DataSetIterator or DataSetIteratorFactory, any other type will throw an IllegalArgumentException
static org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator	getMultiIterator(Object o) Get a DataSetIterator from the given object whether it's a DataSetIterator or DataSetIteratorFactory, any other type will throw an IllegalArgumentException
static double	getScoreFromRegressionEval(org.deeplearning4j.eval.RegressionEvaluation eval, RegressionValue regressionValue) Deprecated.
static double	score(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData, boolean average) Score based on the loss function
static double	score(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testSet, RegressionValue regressionValue) Run a RegressionEvaluation over a DataSetIterator
static double	score(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testData, boolean average) Score based on the loss function
static double	score(org.deeplearning4j.nn.graph.ComputationGraph model, org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testSet, RegressionValue regressionValue)
static double	score(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData, boolean average) Score the given test data with the given multi layer network
static double	score(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model, org.nd4j.linalg.dataset.api.iterator.DataSetIterator testSet, RegressionValue regressionValue) Score the given multi layer network

Methods inherited from class java.lang.Object

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

Constructor Detail

ScoreUtil

public ScoreUtil()

Method Detail

getMultiIterator

public static org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator getMultiIterator(Object o)

Get a DataSetIterator from the given object whether it's a DataSetIterator or DataSetIteratorFactory, any other type will throw an IllegalArgumentException

Parameters:

o - the object to get the iterator from

Returns:

the datasetiterator from the given objects

getIterator

public static org.nd4j.linalg.dataset.api.iterator.DataSetIterator getIterator(Object o)

Get a DataSetIterator from the given object whether it's a DataSetIterator or DataSetIteratorFactory, any other type will throw an IllegalArgumentException

Parameters:

o - the object to get the iterator from

Returns:

the datasetiterator from the given objects

getEvaluation

public static org.deeplearning4j.eval.Evaluation getEvaluation(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model,
                                                               org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData)

Parameters:

model -

testData -

Returns:

getEvaluation

public static org.deeplearning4j.eval.Evaluation getEvaluation(org.deeplearning4j.nn.graph.ComputationGraph model,
                                                               org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testData)

Get the evaluation for the given model and test dataset

Parameters:

model - the model to get the evaluation from

testData - the test data to do the evaluation on

Returns:

the evaluation object with accumulated statistics for the current test data

getEvaluation

public static org.deeplearning4j.eval.Evaluation getEvaluation(org.deeplearning4j.nn.graph.ComputationGraph model,
                                                               org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData)

Get the evaluation for the given model and test dataset

Parameters:

model - the model to get the evaluation from

testData - the test data to do the evaluation on

Returns:

the evaluation object with accumulated statistics for the current test data

score

public static double score(org.deeplearning4j.nn.graph.ComputationGraph model,
                           org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testData,
                           boolean average)

Score based on the loss function

Parameters:

model - the model to score with

testData - the test data to score

average - whether to average the score for the whole batch or not

Returns:

the score for the given test set

score

public static double score(org.deeplearning4j.nn.graph.ComputationGraph model,
                           org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData,
                           boolean average)

Score based on the loss function

Parameters:

model - the model to score with

testData - the test data to score

average - whether to average the score for the whole batch or not

Returns:

the score for the given test set

score

public static double score(org.deeplearning4j.nn.graph.ComputationGraph model,
                           org.nd4j.linalg.dataset.api.iterator.MultiDataSetIterator testSet,
                           RegressionValue regressionValue)

Parameters:

model -

testSet -

regressionValue -

Returns:

score

public static double score(org.deeplearning4j.nn.graph.ComputationGraph model,
                           org.nd4j.linalg.dataset.api.iterator.DataSetIterator testSet,
                           RegressionValue regressionValue)

Run a RegressionEvaluation over a DataSetIterator

Parameters:

model - the model to use

testSet - the test set iterator

regressionValue - the regression type to use

Returns:

score

public static double score(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model,
                           org.nd4j.linalg.dataset.api.iterator.DataSetIterator testData,
                           boolean average)

Score the given test data with the given multi layer network

Parameters:

model - model to use

testData - the test data to test with

average - whether to average the score or not

Returns:

the score for the given test data given the model

score

public static double score(org.deeplearning4j.nn.multilayer.MultiLayerNetwork model,
                           org.nd4j.linalg.dataset.api.iterator.DataSetIterator testSet,
                           RegressionValue regressionValue)

Score the given multi layer network

Parameters:

model - the model to score

testSet - the test set

regressionValue - the regression function to use

Returns:

the score from the given test set

getScoreFromRegressionEval

@Deprecated
public static double getScoreFromRegressionEval(org.deeplearning4j.eval.RegressionEvaluation eval,
                                                            RegressionValue regressionValue)

Deprecated.