org.deeplearning4j.nn.graph.vertex.impl.rnn

Class LastTimeStepVertex

java.lang.Object

org.deeplearning4j.nn.graph.vertex.BaseGraphVertex

org.deeplearning4j.nn.graph.vertex.impl.rnn.LastTimeStepVertex

All Implemented Interfaces:

Serializable, Trainable, GraphVertex

public class LastTimeStepVertex
extends BaseGraphVertex

LastTimeStepVertex is used in the context of recurrent neural network activations, to go from 3d (time series) activations to 2d activations, by extracting out the last time step of activations for each example.
This can be used for example in sequence to sequence architectures, and potentially for sequence classification. NOTE: Because RNNs may have masking arrays (to allow for examples/time series of different lengths in the same minibatch), it is necessary to provide the same of the network input that has the corresponding mask array. If this input does not have a mask array, the last time step of the input will be used for all examples; otherwise, the time step of the last non-zero entry in the mask array (for each example separately) will be used.

Author:

Alex Black

See Also:

Serialized Form

Field Summary

Fields inherited from class org.deeplearning4j.nn.graph.vertex.BaseGraphVertex

epsilon, graph, inputs, inputVertices, outputVertex, outputVertices, vertexIndex, vertexName

Constructor Summary

Constructors

Constructor and Description
LastTimeStepVertex(ComputationGraph graph, String name, int vertexIndex, String inputName)
LastTimeStepVertex(ComputationGraph graph, String name, int vertexIndex, VertexIndices[] inputVertices, VertexIndices[] outputVertices, String inputName)

Method Summary

Modifier and Type	Method and Description
Pair<Gradient,INDArray[]>	doBackward(boolean tbptt, LayerWorkspaceMgr workspaceMgr) Do backward pass
INDArray	doForward(boolean training, LayerWorkspaceMgr workspaceMgr) Do forward pass using the stored inputs
Pair<INDArray,MaskState>	feedForwardMaskArrays(INDArray[] maskArrays, MaskState currentMaskState, int minibatchSize)
Layer	getLayer() Get the Layer (if any).
boolean	hasLayer() Whether the GraphVertex contains a Layer object or not
void	setBackpropGradientsViewArray(INDArray backpropGradientsViewArray) See Model.setBackpropGradientsViewArray(INDArray)
String	toString()

Methods inherited from class org.deeplearning4j.nn.graph.vertex.BaseGraphVertex

canDoBackward, canDoForward, clear, clearVertex, getConfig, getEpsilon, getGradientsViewArray, getInputVertices, getNumInputArrays, getNumOutputConnections, getOutputVertices, getVertexIndex, getVertexName, isInputVertex, numParams, params, paramTable, setEpsilon, setInput, setInputVertices, setLayerAsFrozen, setOutputVertices

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.deeplearning4j.nn.graph.vertex.GraphVertex

getInputs, isOutputVertex, setInputs, setOutputVertex

Constructor Detail

LastTimeStepVertex

public LastTimeStepVertex(ComputationGraph graph,
                          String name,
                          int vertexIndex,
                          String inputName)

LastTimeStepVertex

public LastTimeStepVertex(ComputationGraph graph,
                          String name,
                          int vertexIndex,
                          VertexIndices[] inputVertices,
                          VertexIndices[] outputVertices,
                          String inputName)

Method Detail

hasLayer

public boolean hasLayer()

Description copied from interface: GraphVertex

Whether the GraphVertex contains a Layer object or not

getLayer

public Layer getLayer()

Description copied from interface: GraphVertex

Get the Layer (if any). Returns null if GraphVertex.hasLayer() == false

doForward

public INDArray doForward(boolean training,
                          LayerWorkspaceMgr workspaceMgr)

Description copied from interface: GraphVertex

Do forward pass using the stored inputs

Parameters:

training - if true: forward pass at training time. If false: forward pass at test time

Returns:

The output (for example, activations) of the GraphVertex

doBackward

public Pair<Gradient,INDArray[]> doBackward(boolean tbptt,
                                            LayerWorkspaceMgr workspaceMgr)

Description copied from interface: GraphVertex

Do backward pass

Parameters:

tbptt - If true: do backprop using truncated BPTT

Returns:

The gradients (may be null), and the errors/epsilons for all inputs to this GraphVertex

setBackpropGradientsViewArray

public void setBackpropGradientsViewArray(INDArray backpropGradientsViewArray)

Description copied from interface: GraphVertex

See Model.setBackpropGradientsViewArray(INDArray)

feedForwardMaskArrays

public Pair<INDArray,MaskState> feedForwardMaskArrays(INDArray[] maskArrays,
                                                      MaskState currentMaskState,
                                                      int minibatchSize)

toString

public String toString()

Specified by:

toString in class BaseGraphVertex