Package org.deeplearning4j.nn.conf.layers

Class RecurrentAttentionLayer.Builder

java.lang.Object

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

org.deeplearning4j.nn.conf.layers.samediff.AbstractSameDiffLayer.Builder<T>

org.deeplearning4j.nn.conf.layers.samediff.SameDiffLayer.Builder<RecurrentAttentionLayer.Builder>

org.deeplearning4j.nn.conf.layers.RecurrentAttentionLayer.Builder

Enclosing class:

RecurrentAttentionLayer

public static class RecurrentAttentionLayer.Builder
extends SameDiffLayer.Builder<RecurrentAttentionLayer.Builder>

Field Summary

Fields inherited from class org.deeplearning4j.nn.conf.layers.samediff.SameDiffLayer.Builder

paramWeightInit, weightInit

Fields inherited from class org.deeplearning4j.nn.conf.layers.samediff.AbstractSameDiffLayer.Builder

biasUpdater, regularization, regularizationBias, updater

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

allParamConstraints, biasConstraints, iDropout, layerName, weightConstraints

Constructor Summary

Constructors

Constructor	Description
Builder()

Method Summary

Modifier and Type	Method	Description
RecurrentAttentionLayer.Builder	activation(Activation activation)
RecurrentAttentionLayer	build()
RecurrentAttentionLayer.Builder	hasBias(boolean hasBias)
RecurrentAttentionLayer.Builder	headSize(int headSize)	Size of attention heads
RecurrentAttentionLayer.Builder	nHeads(int nHeads)	Number of Attention Heads
RecurrentAttentionLayer.Builder	nIn(int nIn)
RecurrentAttentionLayer.Builder	nOut(int nOut)
RecurrentAttentionLayer.Builder	projectInput(boolean projectInput)	Project input before applying attention or not.

Methods inherited from class org.deeplearning4j.nn.conf.layers.samediff.SameDiffLayer.Builder

weightInit, weightInit

Methods inherited from class org.deeplearning4j.nn.conf.layers.samediff.AbstractSameDiffLayer.Builder

biasUpdater, l1, l1Bias, l2, l2Bias, regularization, regularizationBias, updater, weightDecay, weightDecay, weightDecayBias, weightDecayBias

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

constrainAllParameters, constrainBias, constrainWeights, dropOut, dropOut, name

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

public Builder()

Method Detail

nIn

public RecurrentAttentionLayer.Builder nIn(int nIn)

Parameters:

nIn - Number of inputs to the layer (input size)

nOut

public RecurrentAttentionLayer.Builder nOut(int nOut)

Parameters:

nOut - Number of outputs (output size)

nHeads

public RecurrentAttentionLayer.Builder nHeads(int nHeads)

Number of Attention Heads

headSize

public RecurrentAttentionLayer.Builder headSize(int headSize)

Size of attention heads

projectInput

public RecurrentAttentionLayer.Builder projectInput(boolean projectInput)

Project input before applying attention or not.

hasBias

public RecurrentAttentionLayer.Builder hasBias(boolean hasBias)

Parameters:

hasBias - If true (default is true) the layer will have a bias

activation

public RecurrentAttentionLayer.Builder activation(Activation activation)

Parameters:

activation - Activation function for the layer

build

public RecurrentAttentionLayer build()

Specified by:

build in class Layer.Builder<RecurrentAttentionLayer.Builder>