weka.dl4j.inference

Class Dl4jCNNExplorer

java.lang.Object

weka.dl4j.inference.Dl4jCNNExplorer

All Implemented Interfaces:

java.io.Serializable, weka.core.CommandlineRunnable, weka.core.OptionHandler

public class Dl4jCNNExplorer
extends java.lang.Object
implements java.io.Serializable, weka.core.OptionHandler, weka.core.CommandlineRunnable

Tool to allow easy experimentation and exploration of a trained ComputationGraph - either from a previously trained Dl4jMlpClassifier, or from a pretrained Zoo Model (default).

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
Dl4jCNNExplorer()

Method Summary

Modifier and Type	Method and Description
void	checkArgs() Check the arguments for the explorer.
void	finishProgress() Close up all progress managers when we finish processing.
void	generateAndSaveOutputMap() Generate the output heatmap, and save to file.
java.awt.image.BufferedImage	generateOutputMap() Generate and return the heatmap.
TopNPredictions	getCurrentPredictions()
CustomModelSetup	getCustomModelSetup()
boolean	getGenerateSaliencyMap()
java.lang.String	getModelName() Get the name of the loaded model
ModelOutputDecoder	getModelOutputDecoder()
java.lang.String[]	getOptions() Gets the current settings of the Classifier.
AbstractCNNSaliencyMapWrapper	getSaliencyMapWrapper()
boolean	getUseCustomModel()
AbstractZooModel	getZooModelType()
void	init() Initialize the ComputationGraph.
java.util.Enumeration<weka.core.Option>	listOptions() Returns an enumeration describing the available options.
void	postExecution() Perform any teardown stuff that might need to happen after execution.
void	preExecution() Perform any setup stuff that might need to happen before execution.
void	processImage(java.io.File imageFile) Performs prediction and (optionally) computes a saliency map.
void	run(java.lang.Object toRun, java.lang.String[] options) Execute the supplied object.
void	setCustomModelSetup(CustomModelSetup customModelSetup)
void	setGenerateSaliencyMap(boolean generateSaliencyMap)
void	setModelOutputDecoder(ModelOutputDecoder modelOutputDecoder)
void	setOptions(java.lang.String[] options) Parses a given list of options.
void	setSaliencyMapWrapper(AbstractCNNSaliencyMapWrapper saliencyMapWrapper)
void	setUseCustomModel(boolean useCustomModel)
void	setZooModelType(AbstractZooModel zooModelType)

Methods inherited from class java.lang.Object

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

Methods inherited from interface weka.core.OptionHandler

makeCopy

Constructor Detail

Dl4jCNNExplorer

public Dl4jCNNExplorer()

Method Detail

init

public void init()
          throws java.lang.Exception

Initialize the ComputationGraph.

Throws:

java.lang.Exception - Exceptions from loading the ComputationGraph

checkArgs

public void checkArgs()
               throws weka.core.WekaException

Check the arguments for the explorer.

Throws:

weka.core.WekaException - If the user hasn't selected a model file

processImage

public void processImage(java.io.File imageFile)
                  throws java.lang.Exception

Performs prediction and (optionally) computes a saliency map.

Parameters:

imageFile - Image to compute.

Throws:

java.lang.Exception - Exception during prediction.

generateOutputMap

public java.awt.image.BufferedImage generateOutputMap()

Generate and return the heatmap.

Returns:

Generated heatmap.

generateAndSaveOutputMap

public void generateAndSaveOutputMap()

Generate the output heatmap, and save to file.

getModelName

public java.lang.String getModelName()

Get the name of the loaded model

Returns:

Model name

getCurrentPredictions

public TopNPredictions getCurrentPredictions()

finishProgress

public void finishProgress()

Close up all progress managers when we finish processing.

getUseCustomModel

@OptionMetadata(commandLineParamIsFlag=true,
                commandLineParamName="useCustomModel",
                commandLineParamSynopsis="-useCustomModel",
                displayName="Use custom-trained model file",
                description="Use the a custom-trained Dl4jMlpClassifier serialized model file, instead of the zoo model.",
                displayOrder=0)
public boolean getUseCustomModel()

setUseCustomModel

public void setUseCustomModel(boolean useCustomModel)

getCustomModelSetup

@OptionMetadata(displayName="Custom model setup",
                description="Options to set if using a custom-trained model",
                commandLineParamName="custom-model",
                commandLineParamSynopsis="-customModel <options>",
                displayOrder=1)
public CustomModelSetup getCustomModelSetup()

setCustomModelSetup

public void setCustomModelSetup(CustomModelSetup customModelSetup)

getZooModelType

@OptionMetadata(displayName="Pretrained zoo model",
                description="Type of pretrained model to use for prediction (instead of trained Dl4jMlpClassifier)",
                commandLineParamName="zooModel",
                commandLineParamSynopsis="-zooModel <options>",
                displayOrder=2)
public AbstractZooModel getZooModelType()

setZooModelType

public void setZooModelType(AbstractZooModel zooModelType)

getModelOutputDecoder

@OptionMetadata(displayName="Model output decoder",
                description="Handles decoding of the model predictions",
                commandLineParamName="decoder",
                commandLineParamSynopsis="-decoder <options>",
                displayOrder=3)
public ModelOutputDecoder getModelOutputDecoder()

setModelOutputDecoder

public void setModelOutputDecoder(ModelOutputDecoder modelOutputDecoder)

getGenerateSaliencyMap

@OptionMetadata(displayName="Generate saliency map",
                description="Should the model explorer generate a ScoreCAM saliency map?",
                commandLineParamName="generate-map",
                commandLineParamSynopsis="-generate-map",
                commandLineParamIsFlag=true,
                displayOrder=4)
public boolean getGenerateSaliencyMap()

setGenerateSaliencyMap

public void setGenerateSaliencyMap(boolean generateSaliencyMap)

getSaliencyMapWrapper

@OptionMetadata(displayName="Saliency map generator",
                description="Saliency map options",
                commandLineParamName="saliency-map",
                commandLineParamSynopsis="-saliency-map <options>",
                displayOrder=5)
public AbstractCNNSaliencyMapWrapper getSaliencyMapWrapper()

setSaliencyMapWrapper

public void setSaliencyMapWrapper(AbstractCNNSaliencyMapWrapper saliencyMapWrapper)

preExecution

public void preExecution()
                  throws java.lang.Exception

Perform any setup stuff that might need to happen before execution.

Specified by:

preExecution in interface weka.core.CommandlineRunnable

Throws:

java.lang.Exception - if a problem occurs during setup

run

public void run(java.lang.Object toRun,
                java.lang.String[] options)
         throws java.lang.Exception

Execute the supplied object.

Specified by:

run in interface weka.core.CommandlineRunnable

Parameters:

toRun - the object to execute

options - any options to pass to the object

Throws:

java.lang.Exception - if a problem occurs.

postExecution

public void postExecution()
                   throws java.lang.Exception

Perform any teardown stuff that might need to happen after execution.

Specified by:

postExecution in interface weka.core.CommandlineRunnable

Throws:

java.lang.Exception - if a problem occurs during teardown

listOptions

public java.util.Enumeration<weka.core.Option> listOptions()

Returns an enumeration describing the available options.

Specified by:

listOptions in interface weka.core.OptionHandler

Returns:

an enumeration of all the available options.

getOptions

public java.lang.String[] getOptions()

Gets the current settings of the Classifier.

Specified by:

getOptions in interface weka.core.OptionHandler

Returns:

an array of strings suitable for passing to setOptions

setOptions

public void setOptions(java.lang.String[] options)
                throws java.lang.Exception

Parses a given list of options.

Specified by:

setOptions in interface weka.core.OptionHandler

Parameters:

options - the list of options as an array of strings

Throws:

java.lang.Exception - if an option is not supported