hex.deeplearning

Class DeepLearningModel

java.lang.Object

water.Iced<T>

water.Keyed<T>

water.Lockable<M>

hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

hex.deeplearning.DeepLearningModel

All Implemented Interfaces:

hex.Model.Contributions, hex.Model.DeepFeatures, java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.api.StreamWriter, water.Freezable<DeepLearningModel>

public class DeepLearningModel
extends hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>
implements hex.Model.DeepFeatures, hex.Model.Contributions

The Deep Learning model It contains a DeepLearningModelInfo with the most up-to-date model, a scoring history, as well as some helpers to indicate the progress

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DeepLearningModel.DeepLearningModelOutput The Deep Learning model output contains a few extra fields in addition to the metrics in Model.Output 1) Scoring history (raw data) 2) weights/biases (raw data) 3) variable importances (TwoDimTable)
static class	DeepLearningModel.DeepLearningParameters Deep Learning Parameters

Nested classes/interfaces inherited from class hex.Model

hex.Model.AdaptFrameParameters, hex.Model.BigScore, hex.Model.BigScoreChunkPredict, hex.Model.BigScorePredict, hex.Model.Contributions, hex.Model.DeepFeatures, hex.Model.ExemplarMembers, hex.Model.FeatureFrequencies, hex.Model.GetMostImportantFeatures, hex.Model.GetNTrees, hex.Model.GLRMArchetypes, hex.Model.GridSortBy, hex.Model.H2OModelDescriptor, hex.Model.InteractionBuilder, hex.Model.InteractionPair, hex.Model.InteractionSpec, hex.Model.JavaModelStreamWriter, hex.Model.JavaScoringOptions, hex.Model.LeafNodeAssignment, hex.Model.Output, hex.Model.Parameters, hex.Model.PredictScoreResult, hex.Model.RowToTreeAssignment, hex.Model.StagedPredictions, hex.Model.UpdateAuxTreeWeights

Nested classes/interfaces inherited from interface hex.Model.Contributions

hex.Model.Contributions.ContributionsOptions, hex.Model.Contributions.ContributionsOutputFormat

Field Summary

Fields

Modifier and Type	Field and Description
long	_timeLastIterationEnter
long	_timeLastScoreStart
water.Key	actual_best_model_key
long	actual_train_samples_per_iteration
double	epoch_counter
int	iterations
water.Key	model_info_key
boolean	stopped_early
double	time_for_communication_us
long	total_checkpointed_run_time_ms
long	total_scoring_time_ms
long	total_setup_time_ms
long	total_training_time_ms
long	training_rows
long	validation_rows

Fields inherited from class hex.Model

_dist, _input_parms, _output, _parms, _toDelete, _warnings, _warningsP, EVAL_AUTO_PARAMS_ENABLED, evalAutoParamsEnabled, scoringInfo

Fields inherited from class water.Lockable

_lockers

Fields inherited from class water.Keyed

_key

Constructor Summary

Constructors

Constructor and Description
DeepLearningModel(water.Key destKey, DeepLearningModel.DeepLearningParameters parms, DeepLearningModel.DeepLearningModelOutput output, water.fvec.Frame train, water.fvec.Frame valid, int nClasses) Regular constructor (from scratch)
DeepLearningModel(water.Key destKey, DeepLearningModel.DeepLearningParameters parms, DeepLearningModel cp, boolean store_best_model, DataInfo dataInfo) Constructor to restart from a checkpointed model

Method Summary

Modifier and Type	Method and Description
protected long	checksum_impl()
DeepLearningModel.DeepLearningParameters	get_params() Get the parameters actually used for model building, not the user-given ones (_parms) They might differ since some defaults are filled in, and some invalid combinations are auto-disabled in modifyParams
protected hex.genmodel.CategoricalEncoding	getGenModelEncoding()
DeepLearningMojoWriter	getMojo()
hex.ToEigenVec	getToEigenVec()
void	initActualParamValues()
boolean	isDistributionHuber()
boolean	isFeatureUsedInPredict(java.lang.String featureName)
DeepLearningScoringInfo	last_scored()
hex.ModelMetrics.MetricBuilder	makeMetricBuilder(java.lang.String[] domain)
double	meanLoss(DataInfo.Row[] myRows) Compute the loss function
DeepLearningModelInfo	model_info()
protected hex.Model.PredictScoreResult	predictScoreImpl(water.fvec.Frame orig, water.fvec.Frame adaptedFr, java.lang.String destination_key, water.Job j, boolean computeMetrics, water.udf.CFuncRef customMetricFunc) Make either a prediction or a reconstruction.
protected water.Futures	remove_impl(water.Futures fs, boolean cascade)
double	score(double[] data)
protected double[]	score0(double[] data, double[] preds)
double[]	score0(double[] data, double[] preds, double offset) Predict from raw double values representing the data
water.fvec.Frame	scoreAutoEncoder(water.fvec.Frame frame, water.Key destination_key, boolean reconstruction_error_per_feature) Score auto-encoded reconstruction (on-the-fly, without allocating the reconstruction as done in Frame score(Frame fr))
water.fvec.Frame	scoreContributions(water.fvec.Frame frame, water.Key<water.fvec.Frame> destination_key, water.Job<water.fvec.Frame> j, hex.Model.Contributions.ContributionsOptions options, water.fvec.Frame backgroundFrame)
water.fvec.Frame	scoreDeepFeatures(water.fvec.Frame frame, int layer) Score auto-encoded reconstruction (on-the-fly, and materialize the deep features of given layer
water.fvec.Frame	scoreDeepFeatures(water.fvec.Frame frame, int layer, water.Job job)
water.fvec.Frame	scoreDeepFeatures(water.fvec.Frame frame, java.lang.String layer, water.Job j)
protected boolean	toJavaCheckTooBig()
protected water.util.SBPrintStream	toJavaInit(water.util.SBPrintStream sb, water.codegen.CodeGeneratorPipeline fileCtx)
protected void	toJavaPredictBody(water.util.SBPrintStream bodySb, water.codegen.CodeGeneratorPipeline classCtx, water.codegen.CodeGeneratorPipeline fileCtx, boolean verboseCode)
hex.VarImp	varImp()

Methods inherited from class hex.Model

adaptFrameForScore, adaptTestForJavaScoring, adaptTestForTrain, adaptTestForTrain, adaptTestForTrain, addMetrics, addModelMetrics, addWarning, aic, auc, AUCPR, classification_error, compareTo, computeDeviances, containsResponse, data, defaultThreshold, defaultThreshold, deleteCrossValidationFoldAssignment, deleteCrossValidationModels, deleteCrossValidationPreds, deviance, deviance, evaluateAutoModelParameters, exportBinaryModel, exportMojo, fetchAll, fillScoringInfo, getDefaultGridSortBy, getPojoInterfaces, haveMojo, havePojo, importBinaryModel, isFeatureUsedInPredict, isGeneric, isSupervised, lift_top_group, likelihood, logloss, loss, mae, makeAdaptFrameParameters, makeBigScoreTask, makeInteraction, makeInteractions, makeInteractions, makePojoWriter, makeSchema, makeScoringDomains, makeScoringNames, makeScoringNames, mean_per_class_error, modelDescriptor, mse, needsPostProcess, postProcessPredictions, r2, readAll_impl, resetThreshold, result, rmsle, score, score, score, score, score, score, score, score0, score0, score0PostProcessSupervised, scoreMetrics, scoring_history, scoringDomains, setInputParms, setupBigScorePredict, testJavaScoring, testJavaScoring, testJavaScoring, testJavaScoring, testJavaScoring, toJava, toJava, toJava, toJavaAlgo, toJavaModelClassName, toJavaTransform, toJavaUUID, toMojo, toMojo, toString, transform, uploadBinaryModel, writeAll_impl, writeTo

Methods inherited from class water.Lockable

delete_and_lock, delete_and_lock, delete_and_lock, delete_and_lock, delete, delete, delete, delete, read_lock, read_lock, read_lock, unlock_all, unlock, unlock, unlock, unlock, update, update, update, write_lock_to_read_lock, write_lock, write_lock, write_lock

Methods inherited from class water.Keyed

checksum_impl, checksum, checksum, getKey, readAll, remove_impl, remove_self_key_impl, remove, remove, remove, remove, remove, remove, removeQuietly, writeAll

Methods inherited from class water.Iced

asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object

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

Methods inherited from interface hex.Model.Contributions

composeScoreContributionTaskMetadata, scoreContributions, scoreContributions, scoreContributions, scoreContributionsWorkEstimate

Field Detail

total_checkpointed_run_time_ms

public long total_checkpointed_run_time_ms

total_training_time_ms

public long total_training_time_ms

total_scoring_time_ms

public long total_scoring_time_ms

total_setup_time_ms

public long total_setup_time_ms

actual_train_samples_per_iteration

public long actual_train_samples_per_iteration

time_for_communication_us

public double time_for_communication_us

epoch_counter

public double epoch_counter

iterations

public int iterations

stopped_early

public boolean stopped_early

training_rows

public long training_rows

validation_rows

public long validation_rows

actual_best_model_key

public water.Key actual_best_model_key

model_info_key

public water.Key model_info_key

_timeLastIterationEnter

public long _timeLastIterationEnter

_timeLastScoreStart

public long _timeLastScoreStart

Constructor Detail

DeepLearningModel

public DeepLearningModel(water.Key destKey,
                         DeepLearningModel.DeepLearningParameters parms,
                         DeepLearningModel cp,
                         boolean store_best_model,
                         DataInfo dataInfo)

Constructor to restart from a checkpointed model

Parameters:

destKey - New destination key for the model

parms - User-given parameters for checkpoint restart

cp - Checkpoint to restart from

store_best_model - Store only the best model instead of the latest one

DeepLearningModel

public DeepLearningModel(water.Key destKey,
                         DeepLearningModel.DeepLearningParameters parms,
                         DeepLearningModel.DeepLearningModelOutput output,
                         water.fvec.Frame train,
                         water.fvec.Frame valid,
                         int nClasses)

Regular constructor (from scratch)

Parameters:

destKey - destination key

parms - DL parameters

output - DL model output

train - Training frame

valid - Validation frame

nClasses - Number of classes (1 for regression or autoencoder)

Method Detail

getToEigenVec

public hex.ToEigenVec getToEigenVec()

Overrides:

getToEigenVec in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

scoreContributions

public water.fvec.Frame scoreContributions(water.fvec.Frame frame,
                                           water.Key<water.fvec.Frame> destination_key,
                                           water.Job<water.fvec.Frame> j,
                                           hex.Model.Contributions.ContributionsOptions options,
                                           water.fvec.Frame backgroundFrame)

Specified by:

scoreContributions in interface hex.Model.Contributions

initActualParamValues

public void initActualParamValues()

Overrides:

initActualParamValues in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

model_info

public final DeepLearningModelInfo model_info()

varImp

public final hex.VarImp varImp()

last_scored

public DeepLearningScoringInfo last_scored()

Overrides:

last_scored in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

get_params

public final DeepLearningModel.DeepLearningParameters get_params()

Get the parameters actually used for model building, not the user-given ones (_parms) They might differ since some defaults are filled in, and some invalid combinations are auto-disabled in modifyParams

Returns:

actually used parameters

makeMetricBuilder

public hex.ModelMetrics.MetricBuilder makeMetricBuilder(java.lang.String[] domain)

Specified by:

makeMetricBuilder in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

predictScoreImpl

protected hex.Model.PredictScoreResult predictScoreImpl(water.fvec.Frame orig,
                                                        water.fvec.Frame adaptedFr,
                                                        java.lang.String destination_key,
                                                        water.Job j,
                                                        boolean computeMetrics,
                                                        water.udf.CFuncRef customMetricFunc)

Make either a prediction or a reconstruction.

Overrides:

predictScoreImpl in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

Parameters:

orig - Test dataset

adaptedFr - Test dataset, adapted to the model

computeMetrics -

Returns:

A frame containing the prediction or reconstruction

score0

protected double[] score0(double[] data,
                          double[] preds)

Specified by:

score0 in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

meanLoss

public double meanLoss(DataInfo.Row[] myRows)

Compute the loss function

Parameters:

myRows - Mini-Batch Array of denseRow's containing numerical/categorical predictor and response data (standardized)

Returns:

loss

score0

public double[] score0(double[] data,
                       double[] preds,
                       double offset)

Predict from raw double values representing the data

Overrides:

score0 in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

Parameters:

data - raw array containing categorical values (horizontalized to 1,0,0,1,0,0 etc.) and numerical values (0.35,1.24,5.3234,etc), both can contain NaNs

preds - predicted label and per-class probabilities (for classification), predicted target (regression), can contain NaNs

Returns:

preds, can contain NaNs

scoreAutoEncoder

public water.fvec.Frame scoreAutoEncoder(water.fvec.Frame frame,
                                         water.Key destination_key,
                                         boolean reconstruction_error_per_feature)

Score auto-encoded reconstruction (on-the-fly, without allocating the reconstruction as done in Frame score(Frame fr))

Specified by:

scoreAutoEncoder in interface hex.Model.DeepFeatures

Parameters:

frame - Original data (can contain response, will be ignored)

destination_key - Frame Id for output

reconstruction_error_per_feature - whether to return the squared error per feature

Returns:

Frame containing one Vec with reconstruction error (MSE) of each reconstructed row, caller is responsible for deletion

scoreDeepFeatures

public water.fvec.Frame scoreDeepFeatures(water.fvec.Frame frame,
                                          int layer)

Score auto-encoded reconstruction (on-the-fly, and materialize the deep features of given layer

Specified by:

scoreDeepFeatures in interface hex.Model.DeepFeatures

Parameters:

frame - Original data (can contain response, will be ignored)

layer - index of the hidden layer for which to extract the features

Returns:

Frame containing the deep features (#cols = hidden[layer])

scoreDeepFeatures

public water.fvec.Frame scoreDeepFeatures(water.fvec.Frame frame,
                                          int layer,
                                          water.Job job)

Specified by:

scoreDeepFeatures in interface hex.Model.DeepFeatures

scoreDeepFeatures

public water.fvec.Frame scoreDeepFeatures(water.fvec.Frame frame,
                                          java.lang.String layer,
                                          water.Job j)

Specified by:

scoreDeepFeatures in interface hex.Model.DeepFeatures

remove_impl

protected water.Futures remove_impl(water.Futures fs,
                                    boolean cascade)

Overrides:

remove_impl in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

toJavaInit

protected water.util.SBPrintStream toJavaInit(water.util.SBPrintStream sb,
                                              water.codegen.CodeGeneratorPipeline fileCtx)

Overrides:

toJavaInit in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

toJavaCheckTooBig

protected boolean toJavaCheckTooBig()

Overrides:

toJavaCheckTooBig in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

toJavaPredictBody

protected void toJavaPredictBody(water.util.SBPrintStream bodySb,
                                 water.codegen.CodeGeneratorPipeline classCtx,
                                 water.codegen.CodeGeneratorPipeline fileCtx,
                                 boolean verboseCode)

Overrides:

toJavaPredictBody in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

checksum_impl

protected long checksum_impl()

Overrides:

checksum_impl in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

getMojo

public DeepLearningMojoWriter getMojo()

Overrides:

getMojo in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

isFeatureUsedInPredict

public boolean isFeatureUsedInPredict(java.lang.String featureName)

Overrides:

isFeatureUsedInPredict in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

isDistributionHuber

public boolean isDistributionHuber()

Overrides:

isDistributionHuber in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

getGenModelEncoding

protected hex.genmodel.CategoricalEncoding getGenModelEncoding()

Overrides:

getGenModelEncoding in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>

score

public double score(double[] data)

Overrides:

score in class hex.Model<DeepLearningModel,DeepLearningModel.DeepLearningParameters,DeepLearningModel.DeepLearningModelOutput>