org.deeplearning4j.nn.conf.layers

Class ConvolutionLayer.BaseConvBuilder<T extends ConvolutionLayer.BaseConvBuilder<T>>

java.lang.Object

org.deeplearning4j.nn.conf.layers.Layer.Builder<T>

org.deeplearning4j.nn.conf.layers.BaseLayer.Builder<T>

org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder<T>

org.deeplearning4j.nn.conf.layers.ConvolutionLayer.BaseConvBuilder<T>

Direct Known Subclasses:

Convolution1DLayer.Builder, Convolution3D.Builder, ConvolutionLayer.Builder, Deconvolution2D.Builder, Deconvolution3D.Builder, DepthwiseConvolution2D.Builder, SeparableConvolution2D.Builder

Enclosing class:

ConvolutionLayer

public abstract static class ConvolutionLayer.BaseConvBuilder<T extends ConvolutionLayer.BaseConvBuilder<T>>
extends FeedForwardLayer.Builder<T>

Field Summary

Fields

Modifier and Type	Field and Description
protected int	convolutionDim
protected ConvolutionMode	convolutionMode Set the convolution mode for the Convolution layer.
protected ConvolutionLayer.AlgoMode	cudnnAlgoMode Defaults to "PREFER_FASTEST", but "NO_WORKSPACE" uses less memory.
protected boolean	cudnnAllowFallback When using CuDNN and an error is encountered, should fallback to the non-CuDNN implementatation be allowed? If set to false, an exception in CuDNN will be propagated back to the user.
protected ConvolutionLayer.BwdDataAlgo	cudnnBwdDataAlgo
protected ConvolutionLayer.BwdFilterAlgo	cudnnBwdFilterAlgo
protected ConvolutionLayer.FwdAlgo	cudnnFwdAlgo
protected int[]	dilation Kernel dilation.
protected boolean	hasBias If true (default): include bias parameters in the model.
int[]	kernelSize
protected int[]	padding
protected int[]	stride

Fields inherited from class org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder

nIn, nOut

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

activationFn, biasInit, biasUpdater, gainInit, gradientNormalization, gradientNormalizationThreshold, iupdater, regularization, regularizationBias, weightInitFn, weightNoise

Fields inherited from class org.deeplearning4j.nn.conf.layers.Layer.Builder

allParamConstraints, biasConstraints, iDropout, layerName, weightConstraints

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	BaseConvBuilder()
protected	BaseConvBuilder(int... kernelSize)
protected	BaseConvBuilder(int[] kernelSize, int[] stride)
protected	BaseConvBuilder(int[] kernelSize, int[] stride, int dim)
protected	BaseConvBuilder(int[] kernelSize, int[] stride, int[] padding)
protected	BaseConvBuilder(int[] kernelSize, int[] stride, int[] padding, int dim)
protected	BaseConvBuilder(int[] kernelSize, int[] stride, int[] padding, int[] dilation)
protected	BaseConvBuilder(int[] kernelSize, int[] stride, int[] padding, int[] dilation, int dim)
protected	BaseConvBuilder(int dim, int... kernelSize)

Method Summary

Modifier and Type	Method and Description
protected abstract boolean	allowCausal()
T	convolutionMode(ConvolutionMode convolutionMode) Set the convolution mode for the Convolution layer.
T	cudnnAlgoMode(ConvolutionLayer.AlgoMode cudnnAlgoMode) Defaults to "PREFER_FASTEST", but "NO_WORKSPACE" uses less memory.
T	cudnnAllowFallback(boolean allowFallback) Deprecated. Use helperAllowFallback(boolean)
T	cudnnBwdDataMode(ConvolutionLayer.BwdDataAlgo cudnnBwdDataAlgo)
T	cudnnBwdFilterMode(ConvolutionLayer.BwdFilterAlgo cudnnBwdFilterAlgo)
T	cudnnFwdMode(ConvolutionLayer.FwdAlgo cudnnFwdAlgo)
T	dilation(int... dilation) Kernel dilation.
T	hasBias(boolean hasBias) If true (default): include bias parameters in the model.
T	helperAllowFallback(boolean allowFallback) When using CuDNN or MKLDNN and an error is encountered, should fallback to the non-helper implementation be allowed? If set to false, an exception in the helper will be propagated back to the user.
T	kernelSize(int... kernelSize)
T	padding(int... padding)
protected void	setConvolutionMode(ConvolutionMode convolutionMode)
T	stride(int... stride)

Methods inherited from class org.deeplearning4j.nn.conf.layers.FeedForwardLayer.Builder

nIn, nIn, nOut, nOut, units

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

activation, activation, biasInit, biasUpdater, dist, gainInit, gradientNormalization, gradientNormalizationThreshold, l1, l1Bias, l2, l2Bias, regularization, regularizationBias, updater, updater, weightDecay, weightDecay, weightDecayBias, weightDecayBias, weightInit, weightInit, weightInit, weightNoise

Methods inherited from class org.deeplearning4j.nn.conf.layers.Layer.Builder

build, constrainAllParameters, constrainBias, constrainWeights, dropOut, dropOut, name

Methods inherited from class java.lang.Object

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

Field Detail

convolutionDim

protected int convolutionDim

hasBias

protected boolean hasBias

If true (default): include bias parameters in the model. False: no bias.

convolutionMode

protected ConvolutionMode convolutionMode

Set the convolution mode for the Convolution layer. See ConvolutionMode for more details

dilation

protected int[] dilation

Kernel dilation. Default: {1, 1}, which is standard convolutions. Used for implementing dilated convolutions, which are also known as atrous convolutions.

For more details, see: Yu and Koltun (2014) and Chen et al. (2014), as well as http://deeplearning.net/software/theano/tutorial/conv_arithmetic.html#dilated-convolutions

kernelSize

public int[] kernelSize

stride

protected int[] stride

padding

protected int[] padding

cudnnAlgoMode

protected ConvolutionLayer.AlgoMode cudnnAlgoMode

Defaults to "PREFER_FASTEST", but "NO_WORKSPACE" uses less memory.

cudnnFwdAlgo

protected ConvolutionLayer.FwdAlgo cudnnFwdAlgo

cudnnBwdFilterAlgo

protected ConvolutionLayer.BwdFilterAlgo cudnnBwdFilterAlgo

cudnnBwdDataAlgo

protected ConvolutionLayer.BwdDataAlgo cudnnBwdDataAlgo

cudnnAllowFallback

protected boolean cudnnAllowFallback

When using CuDNN and an error is encountered, should fallback to the non-CuDNN implementatation be allowed? If set to false, an exception in CuDNN will be propagated back to the user. If false, the built-in (non-CuDNN) implementation for ConvolutionLayer will be used

Constructor Detail

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride,
                          int[] padding,
                          int[] dilation,
                          int dim)

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride,
                          int[] padding,
                          int[] dilation)

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride,
                          int[] padding,
                          int dim)

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride,
                          int[] padding)

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride,
                          int dim)

BaseConvBuilder

protected BaseConvBuilder(int[] kernelSize,
                          int[] stride)

BaseConvBuilder

protected BaseConvBuilder(int dim,
                          int... kernelSize)

BaseConvBuilder

protected BaseConvBuilder(int... kernelSize)

BaseConvBuilder

protected BaseConvBuilder()

Method Detail

allowCausal

protected abstract boolean allowCausal()

setConvolutionMode

protected void setConvolutionMode(ConvolutionMode convolutionMode)

hasBias

public T hasBias(boolean hasBias)

If true (default): include bias parameters in the model. False: no bias.

Parameters:

hasBias - If true: include bias parameters in this model

convolutionMode

public T convolutionMode(ConvolutionMode convolutionMode)

Set the convolution mode for the Convolution layer. See ConvolutionMode for more details

Parameters:

convolutionMode - Convolution mode for layer

dilation

public T dilation(int... dilation)

Kernel dilation. Default: {1, 1}, which is standard convolutions. Used for implementing dilated convolutions, which are also known as atrous convolutions.

For more details, see: Yu and Koltun (2014) and Chen et al. (2014), as well as http://deeplearning.net/software/theano/tutorial/conv_arithmetic.html#dilated-convolutions

Parameters:

dilation - Dilation for kernel

kernelSize

public T kernelSize(int... kernelSize)

stride

public T stride(int... stride)

padding

public T padding(int... padding)

cudnnAlgoMode

public T cudnnAlgoMode(ConvolutionLayer.AlgoMode cudnnAlgoMode)

Defaults to "PREFER_FASTEST", but "NO_WORKSPACE" uses less memory.

cudnnFwdMode

public T cudnnFwdMode(ConvolutionLayer.FwdAlgo cudnnFwdAlgo)

cudnnBwdFilterMode

public T cudnnBwdFilterMode(ConvolutionLayer.BwdFilterAlgo cudnnBwdFilterAlgo)

cudnnBwdDataMode

public T cudnnBwdDataMode(ConvolutionLayer.BwdDataAlgo cudnnBwdDataAlgo)

cudnnAllowFallback

@Deprecated
public T cudnnAllowFallback(boolean allowFallback)

Deprecated. Use helperAllowFallback(boolean)

When using CuDNN and an error is encountered, should fallback to the non-CuDNN implementatation be allowed? If set to false, an exception in CuDNN will be propagated back to the user. If true, the built-in (non-CuDNN) implementation for ConvolutionLayer will be used

Parameters:

allowFallback - Whether fallback to non-CuDNN implementation should be used

helperAllowFallback

public T helperAllowFallback(boolean allowFallback)

When using CuDNN or MKLDNN and an error is encountered, should fallback to the non-helper implementation be allowed? If set to false, an exception in the helper will be propagated back to the user. If true, the built-in (non-MKL/CuDNN) implementation for ConvolutionLayer will be used

Parameters:

allowFallback - Whether fallback to non-CuDNN implementation should be used