Package de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers

Class TransformersFineTunerHpSearch

java.lang.Object

eu.sealsproject.platform.res.tool.impl.AbstractPlugin

de.uni_mannheim.informatik.dws.melt.matching_base.MatcherURL

de.uni_mannheim.informatik.dws.melt.matching_base.MatcherFile

de.uni_mannheim.informatik.dws.melt.matching_jena.MatcherYAAA

de.uni_mannheim.informatik.dws.melt.matching_jena.MatcherYAAAJena

de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBase

de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBaseFineTuner

de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersFineTuner

de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersFineTunerHpSearch

All Implemented Interfaces:

Filter, IMatcher<org.apache.jena.ontology.OntModel,Alignment,Properties>, eu.sealsproject.platform.res.domain.omt.IOntologyMatchingToolBridge, eu.sealsproject.platform.res.tool.api.IPlugin, eu.sealsproject.platform.res.tool.api.IToolBridge

public class TransformersFineTunerHpSearch
extends TransformersFineTuner
implements Filter

Field Summary

Fields

Modifier and Type	Field	Description
private TransformersHpSearchSpace	hpMutations	These are the mutations for the hyperparameter - only applicable for PBT (population based training).
private TransformersHpSearchSpace	hpSpace	The initial hyper parameter search space.
private static final org.slf4j.Logger	LOGGER
private int	numberOfTrials	the number of different hyperparamter combinations which are tried out
private TransformersOptimizingMetric	optimizingMetric	The metric to optimize during hyperparameter search.
private float	testSize	A number between zero and one which represents the proportion of the data to include in the test split

Fields inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersFineTuner

batchSizeOptimization

Fields inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBaseFineTuner

additionallySwitchSourceTarget, NEWLINE, resultingModelLocation, trainingFile

Fields inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBase

cudaVisibleDevices, extractor, modelName, multipleTextsToMultipleExamples, multiProcessing, trainingArguments, transformersCache, usingTensorflow

Fields inherited from class de.uni_mannheim.informatik.dws.melt.matching_base.MatcherFile

FILE_PREFIX, FILE_SUFFIX

Constructor Summary

Constructors

Constructor	Description
TransformersFineTunerHpSearch(TextExtractorMap extractor, String initialModelName, File resultingModelLocation)
TransformersFineTunerHpSearch(TextExtractor extractor, String initialModelName, File resultingModelLocation)

Method Summary

Modifier and Type	Method	Description
File	finetuneModel(File trainingFile)	Finetune a given model with the provided text in the csv file (three columns: first text, second text, label(0/1))
TransformersHpSearchSpace	getHpMutations()	Returns the mutations for the hyperparameter - only applicable for PBT (population based training).
TransformersHpSearchSpace	getHpSpace()	Returns the initial hyper parameter search space.
int	getNumberOfTrials()	Returns the number of trials which should be executed during hyperparameter search.
TransformersOptimizingMetric	getOptimizingMetric()	Returns the metric which is optimized during hyperparameter search.
float	getTestSize()	Returns a number between zero and one which represents the proportion of the data to include in the test split.
void	setHpMutations(TransformersHpSearchSpace hpMutations)	Sets the mutations for the hyperparameter - only applicable for PBT (population based training).
void	setHpSpace(TransformersHpSearchSpace hpSpace)	Sets the initial hyper parameter search space.
void	setNumberOfTrials(int numberOfTrials)	Sets the number of trials which should be executed during hyperparameter search.
void	setOptimizingMetric(TransformersOptimizingMetric optimizingMetric)	Sets the metric which is optimized during hyperparameter search.
void	setTestSize(float testSize)	Sets the number between zero and one which represents the proportion of the data to include in the test split

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersFineTuner

addTrainingParameterToMakeTrainingFaster, getBatchSizeOptimization, getMaximumPerDeviceTrainBatchSize, getMaximumPerDeviceTrainBatchSize, isAdjustMaxBatchSize, setAdjustMaxBatchSize, setBatchSizeOptimization

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBaseFineTuner

clearTrainingData, createTrainingFile, finetuneModel, getResultingModelLocation, getTrainingFile, isAdditionallySwitchSourceTarget, match, setAdditionallySwitchSourceTarget, setResultingModelLocation, writeClassificationFormat, writeTrainingFile

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_ml.python.nlptransformers.TransformersBase

addTrainingArgument, getCudaVisibleDevices, getCudaVisibleDevicesButOnlyOneGPU, getExamplesForBatchSizeOptimization, getExtractor, getExtractorMap, getModelName, getMultiProcessing, getTextualRepresentation, getTrainingArguments, getTransformersCache, isMultipleTextsToMultipleExamples, isOptimizeForMixedPrecisionTraining, isUsingTensorflow, setCudaVisibleDevices, setCudaVisibleDevices, setExtractor, setExtractorMap, setModelName, setMultipleTextsToMultipleExamples, setMultiProcessing, setOptimizeForMixedPrecisionTraining, setTrainingArguments, setTransformersCache, setUsingTensorflow, writeExamplesToFile

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_jena.MatcherYAAAJena

getModelSpec, match, readOntology

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_jena.MatcherYAAA

match

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_base.MatcherFile

match

Methods inherited from class de.uni_mannheim.informatik.dws.melt.matching_base.MatcherURL

align, align, canExecute, getType

Methods inherited from class eu.sealsproject.platform.res.tool.impl.AbstractPlugin

getId, getVersion, setId, setVersion

Methods inherited from class java.lang.Object

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

Methods inherited from interface eu.sealsproject.platform.res.tool.api.IPlugin

getId, getVersion

Field Details

LOGGER

private static final org.slf4j.Logger LOGGER

numberOfTrials

private int numberOfTrials

the number of different hyperparamter combinations which are tried out

testSize

private float testSize

A number between zero and one which represents the proportion of the data to include in the test split

optimizingMetric

private TransformersOptimizingMetric optimizingMetric

The metric to optimize during hyperparameter search.

hpSpace

private TransformersHpSearchSpace hpSpace

The initial hyper parameter search space. For nearly all algorithms except PBT, this is the only search space which should be defined.

hpMutations

private TransformersHpSearchSpace hpMutations

These are the mutations for the hyperparameter - only applicable for PBT (population based training).

Constructor Details

TransformersFineTunerHpSearch

public TransformersFineTunerHpSearch(TextExtractorMap extractor,
 String initialModelName,
 File resultingModelLocation)

TransformersFineTunerHpSearch

public TransformersFineTunerHpSearch(TextExtractor extractor,
 String initialModelName,
 File resultingModelLocation)

Method Details

finetuneModel

public File finetuneModel(File trainingFile)
                   throws Exception

Finetune a given model with the provided text in the csv file (three columns: first text, second text, label(0/1))

Overrides:

finetuneModel in class TransformersFineTuner

Parameters:

trainingFile - csv file with three columns: first text, second text, label(0/1)

Returns:

the final location (directory) of the finetuned model (which is also given in the constructor)

Throws:

Exception - in case of any error

getNumberOfTrials

public int getNumberOfTrials()

Returns the number of trials which should be executed during hyperparameter search. This means how many different hyperparameter combinations should be tried out. The more the better, but also takes more time.

Returns:

the number of different hyperparamter combinations which are executed

setNumberOfTrials

public void setNumberOfTrials(int numberOfTrials)

Sets the number of trials which should be executed during hyperparameter search. This means how many different hyperparameter combinations should be tried out. The more the better, but also takes more time.

Parameters:

numberOfTrials - the number of different hyperparamter combinations which are executed

getTestSize

public float getTestSize()

Returns a number between zero and one which represents the proportion of the data to include in the test split.

Returns:

a number between zero and one which represents the proportion of the data to include in the test split

setTestSize

public void setTestSize(float testSize)

Sets the number between zero and one which represents the proportion of the data to include in the test split

Parameters:

testSize - number between zero and one which represents the proportion of the data to include in the test split

getOptimizingMetric

public TransformersOptimizingMetric getOptimizingMetric()

Returns the metric which is optimized during hyperparameter search.

Returns:

the metric which is optimized during hyperparameter search

setOptimizingMetric

public void setOptimizingMetric(TransformersOptimizingMetric optimizingMetric)

Sets the metric which is optimized during hyperparameter search.

Parameters:

optimizingMetric - the metric which is optimized during hyperparameter search

getHpSpace

public TransformersHpSearchSpace getHpSpace()

Returns the initial hyper parameter search space. For nearly all algorithms except PBT, this is the only search space which should be defined.

Returns:

the initial hyper parameter search space.

setHpSpace

public void setHpSpace(TransformersHpSearchSpace hpSpace)

Sets the initial hyper parameter search space. For nearly all algorithms except PBT, this is the only search space which should be defined.

Parameters:

hpSpace - the initial hyper parameter search space

getHpMutations

public TransformersHpSearchSpace getHpMutations()

Returns the mutations for the hyperparameter - only applicable for PBT (population based training).

Returns:

the mutations for the hyperparameter

setHpMutations

public void setHpMutations(TransformersHpSearchSpace hpMutations)

Sets the mutations for the hyperparameter - only applicable for PBT (population based training).

Parameters:

hpMutations - the mutations for the hyperparameter