org.deeplearning4j.dbn

Class GaussianRectifiedLinearDBN

java.lang.Object

org.deeplearning4j.nn.BaseMultiLayerNetwork

org.deeplearning4j.dbn.DBN

org.deeplearning4j.dbn.GaussianRectifiedLinearDBN

All Implemented Interfaces:

Serializable, Persistable

public class GaussianRectifiedLinearDBN
extends DBN

Rectified Linear Units for activations and gaussian inputs

Author:

Adam Gibson

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	GaussianRectifiedLinearDBN.Builder

Field Summary

Fields inherited from class org.deeplearning4j.nn.BaseMultiLayerNetwork

dropOut, errorTolerance, gradientListeners, learningRateUpdate, lossFunction, multiLayerGradientListeners, normalizeByInputRows, optimizationAlgorithm

Constructor Summary

Constructors

Constructor and Description
GaussianRectifiedLinearDBN()
GaussianRectifiedLinearDBN(int nIns, int[] hiddenLayerSizes, int nOuts, int nLayers, org.apache.commons.math3.random.RandomGenerator rng)
GaussianRectifiedLinearDBN(int nIn, int[] hiddenLayerSizes, int nOuts, int nLayers, org.apache.commons.math3.random.RandomGenerator rng, org.jblas.DoubleMatrix input, org.jblas.DoubleMatrix labels)

Method Summary

Methods

Modifier and Type	Method and Description
HiddenLayer	createHiddenLayer(int index, int nIn, int nOut, ActivationFunction activation, org.apache.commons.math3.random.RandomGenerator rng, org.jblas.DoubleMatrix layerInput, org.apache.commons.math3.distribution.RealDistribution dist) Creates a hidden layer with the given parameters.
NeuralNetwork	createLayer(org.jblas.DoubleMatrix input, int nVisible, int nHidden, org.jblas.DoubleMatrix W, org.jblas.DoubleMatrix hBias, org.jblas.DoubleMatrix vBias, org.apache.commons.math3.random.RandomGenerator rng, int index) Creates a layer depending on the index.
NeuralNetwork[]	createNetworkLayers(int numLayers)

Methods inherited from class org.deeplearning4j.dbn.DBN

pretrain, pretrain, pretrain, trainNetwork

Methods inherited from class org.deeplearning4j.nn.BaseMultiLayerNetwork

applyTransforms, asDecoder, backProp, backPropStep, clone, encode, fanIn, feedForward, feedForward, finetune, finetune, getActivation, getColumnMeans, getColumnStds, getColumnSums, getDist, getDropOut, getErrorTolerance, getFanIn, getGradient, getHiddenBiasTransforms, getHiddenLayerSizes, getInput, getL2, getLabels, getLayers, getLearningRateUpdate, getLogLayer, getLossFunction, getMomentum, getnIns, getnLayers, getnOuts, getOptimizationAlgorithm, getOptimizer, getReconstructionCrossEntropy, getRenderWeightsEveryNEpochs, getRng, getSigmoidLayers, getSparsity, getVisibleBiasTransforms, getWeightTransforms, init, initialize, initializeLayers, initializeNetwork, isForceNumEpochs, isNormalizeByInputRows, isShouldBackProp, isShouldInit, isToDecode, isUseAdaGrad, isUseHiddenActivationsForwardProp, isUseRegularization, load, loadFromFile, merge, negativeLogLikelihood, predict, reconstruct, reconstruct, resetAdaGrad, setActivation, setColumnMeans, setColumnStds, setColumnSums, setDist, setDropOut, setErrorTolerance, setFanIn, setForceNumEpochs, setHiddenLayerSizes, setInput, setL2, setLabels, setLayers, setLearningRateUpdate, setLogLayer, setLossFunction, setMomentum, setnIns, setnLayers, setNormalizeByInputRows, setnOuts, setOptimizationAlgorithm, setOptimizer, setRenderWeightsEveryNEpochs, setRng, setShouldBackProp, setShouldInit, setSigmoidLayers, setSparsity, setToDecode, setUseAdaGrad, setUseHiddenActivationsForwardProp, setUseRegularization, setWeightTransforms, synchonrizeRng, update, write

Methods inherited from class java.lang.Object

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

Constructor Detail

GaussianRectifiedLinearDBN

public GaussianRectifiedLinearDBN()

GaussianRectifiedLinearDBN

public GaussianRectifiedLinearDBN(int nIn,
                          int[] hiddenLayerSizes,
                          int nOuts,
                          int nLayers,
                          org.apache.commons.math3.random.RandomGenerator rng,
                          org.jblas.DoubleMatrix input,
                          org.jblas.DoubleMatrix labels)

GaussianRectifiedLinearDBN

public GaussianRectifiedLinearDBN(int nIns,
                          int[] hiddenLayerSizes,
                          int nOuts,
                          int nLayers,
                          org.apache.commons.math3.random.RandomGenerator rng)

Method Detail

createLayer

public NeuralNetwork createLayer(org.jblas.DoubleMatrix input,
                        int nVisible,
                        int nHidden,
                        org.jblas.DoubleMatrix W,
                        org.jblas.DoubleMatrix hBias,
                        org.jblas.DoubleMatrix vBias,
                        org.apache.commons.math3.random.RandomGenerator rng,
                        int index)

Description copied from class: BaseMultiLayerNetwork

Creates a layer depending on the index. The main reason this matters is for continuous variations such as the CDBN where the first layer needs to be an CRBM for continuous inputs. Please be sure to call super.initializeNetwork to handle the passing of baseline parameters such as fanin and rendering.

Overrides:

createLayer in class DBN

Parameters:

input - the input to the layer

nVisible - the number of visible inputs

nHidden - the number of hidden units

W - the weight vector

hBias - the hidden bias

vBias - the visible bias

rng - the rng to use (THiS IS IMPORTANT; YOU DO NOT WANT TO HAVE A MIS REFERENCED RNG OTHERWISE NUMBERS WILL BE MEANINGLESS)

index - the index of the layer

Returns:

a neural network layer such as RBM

createHiddenLayer

public HiddenLayer createHiddenLayer(int index,
                            int nIn,
                            int nOut,
                            ActivationFunction activation,
                            org.apache.commons.math3.random.RandomGenerator rng,
                            org.jblas.DoubleMatrix layerInput,
                            org.apache.commons.math3.distribution.RealDistribution dist)

Creates a hidden layer with the given parameters. The default implementation is a binomial sampling hidden layer, but this can be overriden for other kinds of hidden units

Overrides:

createHiddenLayer in class BaseMultiLayerNetwork

Parameters:

nIn - the number of inputs

nOut - the number of outputs

activation - the activation function for the layer

rng - the rng to use for sampling

layerInput - the layer starting input

dist - the probability distribution to use for generating weights

Returns:

a hidden layer with the given paremters

createNetworkLayers

public NeuralNetwork[] createNetworkLayers(int numLayers)

Overrides:

createNetworkLayers in class DBN