org.deeplearning4j.nn.layers.convolution

Class ConvolutionLayer

java.lang.Object

org.deeplearning4j.nn.layers.AbstractLayer<LayerConfT>

org.deeplearning4j.nn.layers.BaseLayer<ConvolutionLayer>

org.deeplearning4j.nn.layers.convolution.ConvolutionLayer

All Implemented Interfaces:

Serializable, Cloneable, Layer, Model, Trainable

Direct Known Subclasses:

Convolution1DLayer, Convolution3DLayer, Deconvolution2DLayer, DepthwiseConvolution2DLayer, SeparableConvolution2DLayer

public class ConvolutionLayer
extends BaseLayer<ConvolutionLayer>

Convolution layer

Author:

Adam Gibson (original impl), Alex Black (current version)

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from interface org.deeplearning4j.nn.api.Layer

Layer.TrainingMode, Layer.Type

Field Summary

Fields

Modifier and Type	Field and Description
protected ConvolutionMode	convolutionMode
protected INDArray	dummyBias
protected INDArray	dummyBiasGrad
protected ConvolutionHelper	helper
protected int	helperCountFail
protected INDArray	i2d

Fields inherited from class org.deeplearning4j.nn.layers.BaseLayer

gradient, gradientsFlattened, gradientViews, optimizer, params, paramsFlattened, score, solver, weightNoiseParams

Fields inherited from class org.deeplearning4j.nn.layers.AbstractLayer

cacheMode, conf, dataType, dropoutApplied, epochCount, index, input, inputModificationAllowed, iterationCount, maskArray, maskState, preOutput, trainingListeners

Constructor Summary

Constructors

Constructor and Description
ConvolutionLayer(NeuralNetConfiguration conf, DataType dataType)

Method Summary

Modifier and Type	Method and Description
INDArray	activate(boolean training, LayerWorkspaceMgr workspaceMgr) Perform forward pass and return the activations array with the last set input
Pair<Gradient,INDArray>	backpropGradient(INDArray epsilon, LayerWorkspaceMgr workspaceMgr) Calculate the gradient relative to the error in the next layer
Pair<INDArray,MaskState>	feedForwardMaskArray(INDArray maskArray, MaskState currentMaskState, int minibatchSize) Feed forward the input mask array, setting in the layer as appropriate.
void	fit(INDArray input, LayerWorkspaceMgr workspaceMgr) Fit the model to the given data
LayerHelper	getHelper()
boolean	hasBias() Does this layer have no bias term? Many layers (dense, convolutional, output, embedding) have biases by default, but no-bias versions are possible via configuration
boolean	isPretrainLayer() Returns true if the layer can be trained in an unsupervised/pretrain manner (AE, VAE, etc)
protected Pair<INDArray,INDArray>	preOutput(boolean training, boolean forBackprop, LayerWorkspaceMgr workspaceMgr) PreOutput method that also returns the im2col2d array (if being called for backprop), as this can be re-used instead of being calculated again.
protected Pair<INDArray,INDArray>	preOutput4d(boolean training, boolean forBackprop, LayerWorkspaceMgr workspaceMgr) preOutput4d: Used so that ConvolutionLayer subclasses (such as Convolution1DLayer) can maintain their standard non-4d preOutput method, while overriding this to return 4d activations (for use in backprop) without modifying the public API
void	setParams(INDArray params) Set the parameters for this model.
Layer.Type	type() Returns the layer type
protected void	validateInputDepth(long inDepth)
protected void	validateInputRank()

Methods inherited from class org.deeplearning4j.nn.layers.BaseLayer

calcRegularizationScore, clear, clearNoiseWeightParams, clone, computeGradientAndScore, fit, getGradientsViewArray, getOptimizer, getParam, getParamWithNoise, gradient, hasLayerNorm, layerConf, numParams, params, paramTable, paramTable, preOutput, preOutputWithPreNorm, score, setBackpropGradientsViewArray, setParam, setParams, setParamsViewArray, setParamTable, setScoreWithZ, toString, update, update

Methods inherited from class org.deeplearning4j.nn.layers.AbstractLayer

activate, addListeners, allowInputModification, applyConstraints, applyDropOutIfNecessary, applyMask, assertInputSet, backpropDropOutIfPresent, batchSize, close, conf, getConfig, getEpochCount, getIndex, getInput, getInputMiniBatchSize, getListeners, getMaskArray, gradientAndScore, init, input, layerId, numParams, setCacheMode, setConf, setEpochCount, setIndex, setInput, setInputMiniBatchSize, setListeners, setListeners, setMaskArray, updaterDivideByMinibatch

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.deeplearning4j.nn.api.Layer

getIterationCount, setIterationCount

Field Detail

i2d

protected INDArray i2d

helper

protected ConvolutionHelper helper

helperCountFail

protected int helperCountFail

convolutionMode

protected ConvolutionMode convolutionMode

dummyBias

protected transient INDArray dummyBias

dummyBiasGrad

protected transient INDArray dummyBiasGrad

Constructor Detail

ConvolutionLayer

public ConvolutionLayer(NeuralNetConfiguration conf,
                        DataType dataType)

Method Detail

type

public Layer.Type type()

Description copied from interface: Layer

Returns the layer type

Specified by:

type in interface Layer

Overrides:

type in class AbstractLayer<ConvolutionLayer>

Returns:

backpropGradient

public Pair<Gradient,INDArray> backpropGradient(INDArray epsilon,
                                                LayerWorkspaceMgr workspaceMgr)

Description copied from interface: Layer

Calculate the gradient relative to the error in the next layer

Specified by:

backpropGradient in interface Layer

Overrides:

backpropGradient in class BaseLayer<ConvolutionLayer>

Parameters:

epsilon - w^(L+1)*delta^(L+1). Or, equiv: dC/da, i.e., (dC/dz)*(dz/da) = dC/da, where C is cost function a=sigma(z) is activation.

workspaceMgr - Workspace manager

Returns:

Pair where Gradient is gradient for this layer, INDArray is epsilon (activation gradient) needed by next layer, but before element-wise multiply by sigmaPrime(z). So for standard feed-forward layer, if this layer is L, then return.getSecond() == dL/dIn = (w^(L)*(delta^(L))^T)^T. Note that the returned array should be placed in the ArrayType.ACTIVATION_GRAD workspace via the workspace manager

preOutput4d

protected Pair<INDArray,INDArray> preOutput4d(boolean training,
                                              boolean forBackprop,
                                              LayerWorkspaceMgr workspaceMgr)

preOutput4d: Used so that ConvolutionLayer subclasses (such as Convolution1DLayer) can maintain their standard non-4d preOutput method, while overriding this to return 4d activations (for use in backprop) without modifying the public API

validateInputRank

protected void validateInputRank()

validateInputDepth

protected void validateInputDepth(long inDepth)

preOutput

protected Pair<INDArray,INDArray> preOutput(boolean training,
                                            boolean forBackprop,
                                            LayerWorkspaceMgr workspaceMgr)

PreOutput method that also returns the im2col2d array (if being called for backprop), as this can be re-used instead of being calculated again.

Parameters:

training - Train or test time (impacts dropout)

forBackprop - If true: return the im2col2d array for re-use during backprop. False: return null for second pair entry. Note that it may still be null in the case of CuDNN and the like.

Returns:

Pair of arrays: preOutput (activations) and optionally the im2col2d array

activate

public INDArray activate(boolean training,
                         LayerWorkspaceMgr workspaceMgr)

Description copied from interface: Layer

Perform forward pass and return the activations array with the last set input

Specified by:

activate in interface Layer

Overrides:

activate in class BaseLayer<ConvolutionLayer>

Parameters:

training - training or test mode

workspaceMgr - Workspace manager

Returns:

the activation (layer output) of the last specified input. Note that the returned array should be placed in the ArrayType.ACTIVATIONS workspace via the workspace manager

hasBias

public boolean hasBias()

Description copied from class: BaseLayer

Does this layer have no bias term? Many layers (dense, convolutional, output, embedding) have biases by default, but no-bias versions are possible via configuration

Overrides:

hasBias in class BaseLayer<ConvolutionLayer>

Returns:

True if a bias term is present, false otherwise

isPretrainLayer

public boolean isPretrainLayer()

Description copied from interface: Layer

Returns true if the layer can be trained in an unsupervised/pretrain manner (AE, VAE, etc)

Returns:

true if the layer can be pretrained (using fit(INDArray), false otherwise

getHelper

public LayerHelper getHelper()

Specified by:

getHelper in interface Layer

Overrides:

getHelper in class AbstractLayer<ConvolutionLayer>

Returns:

Get the layer helper, if any

fit

public void fit(INDArray input,
                LayerWorkspaceMgr workspaceMgr)

Description copied from interface: Model

Fit the model to the given data

Specified by:

fit in interface Model

Overrides:

fit in class BaseLayer<ConvolutionLayer>

Parameters:

input - the data to fit the model to

setParams

public void setParams(INDArray params)

Description copied from interface: Model

Set the parameters for this model. This expects a linear ndarray which then be unpacked internally relative to the expected ordering of the model

Specified by:

setParams in interface Model

Overrides:

setParams in class BaseLayer<ConvolutionLayer>

Parameters:

params - the parameters for the model

feedForwardMaskArray

public Pair<INDArray,MaskState> feedForwardMaskArray(INDArray maskArray,
                                                     MaskState currentMaskState,
                                                     int minibatchSize)

Description copied from interface: Layer

Feed forward the input mask array, setting in the layer as appropriate. This allows different layers to handle masks differently - for example, bidirectional RNNs and normal RNNs operate differently with masks (the former sets activations to 0 outside of the data present region (and keeps the mask active for future layers like dense layers), whereas normal RNNs don't zero out the activations/errors )instead relying on backpropagated error arrays to handle the variable length case.
This is also used for example for networks that contain global pooling layers, arbitrary preprocessors, etc.

Specified by:

feedForwardMaskArray in interface Layer

Overrides:

feedForwardMaskArray in class AbstractLayer<ConvolutionLayer>

Parameters:

maskArray - Mask array to set

currentMaskState - Current state of the mask - see MaskState

minibatchSize - Current minibatch size. Needs to be known as it cannot always be inferred from the activations array due to reshaping (such as a DenseLayer within a recurrent neural network)

Returns:

New mask array after this layer, along with the new mask state.