hex.deepwater

Class DeepWaterModel

java.lang.Object

water.Iced<T>

water.Keyed<T>

water.Lockable<M>

hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

hex.deepwater.DeepWaterModel

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable<DeepWaterModel>

public class DeepWaterModel
extends hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

The Deep Learning model It contains a DeepWaterModelInfo 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/interfaces inherited from class hex.Model

hex.Model.BigScore, hex.Model.DeepFeatures, hex.Model.ExemplarMembers, hex.Model.GetMostImportantFeatures, hex.Model.GLRMArchetypes, hex.Model.InteractionPair, hex.Model.JavaModelStreamWriter, hex.Model.LeafNodeAssignment, hex.Model.Output, hex.Model.Parameters

Field Summary

Fields inherited from class hex.Model

_dist, _output, _parms, _toDelete, _warnings, scoringInfo

Fields inherited from class water.Lockable

_lockers

Fields inherited from class water.Keyed

_key

Constructor Summary

Constructors

Constructor and Description
DeepWaterModel(water.Key<DeepWaterModel> destKey, DeepWaterParameters parms, DeepWaterModel cp, DataInfo dataInfo) Constructor to restart from a checkpointed model
DeepWaterModel(water.Key<DeepWaterModel> destKey, DeepWaterParameters params, DeepWaterModelOutput output, water.fvec.Frame train, water.fvec.Frame valid, int nClasses) Regular constructor (from scratch)

Method Summary

Methods

Modifier and Type	Method and Description
protected long	checksum_impl()
DeepWaterParameters	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
DeepwaterMojoWriter	getMojo()
hex.ToEigenVec	getToEigenVec()
DeepWaterScoringInfo	last_scored()
hex.ModelMetrics.MetricBuilder	makeMetricBuilder(java.lang.String[] domain)
DeepWaterModelInfo	model_info()
protected water.fvec.Frame	predictScoreImpl(water.fvec.Frame fr, water.fvec.Frame adaptFrm, java.lang.String destination_key, water.Job j, boolean computeMetrics)
protected water.Futures	remove_impl(water.Futures fs)
water.api.schemas3.ModelSchemaV3	schema()
double[]	score0(water.fvec.Chunk[] chks, double weight, double offset, int row_in_chunk, double[] tmp, double[] preds)
protected double[]	score0(double[] data, double[] preds) Single-instance scoring - slow, not optimized for mini-batches - do not use unless you know what you're doing
double[]	score0(double[] data, double[] preds, double weight, double offset)
protected hex.ModelMetrics.MetricBuilder	scoreMetrics(water.fvec.Frame adaptFrm)

Methods inherited from class hex.Model

adaptTestForTrain, adaptTestForTrain, addMetrics, addModelMetrics, addWarning, auc, classification_error, cleanup_adapt, compareTo, computeDeviances, data, defaultThreshold, deleteCrossValidationModels, deviance, deviance, fillScoringInfo, isSupervised, lift_top_group, logloss, loss, mae, makeBigScoreTask, makeInteraction, makeInteractions, makeInteractions, makeSchema, makeScoringNames, mean_per_class_error, mse, readAll_impl, rmsle, score, score, score, score, score, score0, scoring_history, scoringDomains, testJavaScoring, testJavaScoring, testJavaScoring, toJava, toJava, toJava, toJavaCheckTooBig, toJavaInit, toJavaNCLASSES, toJavaPredictBody, toJavaPROB, toJavaSuper, toString, writeAll_impl

Methods inherited from class water.Lockable

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

Methods inherited from class water.Keyed

checksum, readAll, remove, remove, remove, remove, writeAll

Methods inherited from class water.Iced

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

Methods inherited from class java.lang.Object

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

Constructor Detail

DeepWaterModel

public DeepWaterModel(water.Key<DeepWaterModel> destKey,
              DeepWaterParameters parms,
              DeepWaterModel cp,
              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

DeepWaterModel

public DeepWaterModel(water.Key<DeepWaterModel> destKey,
              DeepWaterParameters params,
              DeepWaterModelOutput output,
              water.fvec.Frame train,
              water.fvec.Frame valid,
              int nClasses)

Regular constructor (from scratch)

Parameters:

destKey - destination key

params - DL parameters

output - DL model output

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

Method Detail

getMojo

public DeepwaterMojoWriter getMojo()

Overrides:

getMojo in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

schema

public water.api.schemas3.ModelSchemaV3 schema()

model_info

public final DeepWaterModelInfo model_info()

getToEigenVec

public hex.ToEigenVec getToEigenVec()

Overrides:

getToEigenVec in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

last_scored

public DeepWaterScoringInfo last_scored()

Overrides:

last_scored in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

get_params

public final DeepWaterParameters 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<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

score0

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

Single-instance scoring - slow, not optimized for mini-batches - do not use unless you know what you're doing

Specified by:

score0 in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

Parameters:

data - One single observation unrolled into a double[], with a length equal to the number of input neurons

preds - Array to store the predictions in (nclasses+1)

Returns:

vector of [0, p0, p1, p2, etc.]

score0

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

Overrides:

score0 in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

score0

public double[] score0(water.fvec.Chunk[] chks,
              double weight,
              double offset,
              int row_in_chunk,
              double[] tmp,
              double[] preds)

Overrides:

score0 in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

checksum_impl

protected long checksum_impl()

Overrides:

checksum_impl in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

predictScoreImpl

protected water.fvec.Frame predictScoreImpl(water.fvec.Frame fr,
                                water.fvec.Frame adaptFrm,
                                java.lang.String destination_key,
                                water.Job j,
                                boolean computeMetrics)

Overrides:

predictScoreImpl in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

scoreMetrics

protected hex.ModelMetrics.MetricBuilder scoreMetrics(water.fvec.Frame adaptFrm)

Overrides:

scoreMetrics in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>

remove_impl

protected water.Futures remove_impl(water.Futures fs)

Overrides:

remove_impl in class hex.Model<DeepWaterModel,DeepWaterParameters,DeepWaterModelOutput>