org.deeplearning4j.nn.conf

Interface InputPreProcessor

All Superinterfaces:

java.lang.Cloneable, java.io.Serializable

All Known Implementing Classes:

BaseInputPreProcessor, BinomialSamplingPreProcessor, CnnToFeedForwardPreProcessor, CnnToRnnPreProcessor, ComposableInputPreProcessor, FeedForwardToCnnPreProcessor, FeedForwardToRnnPreProcessor, MyCustomPreprocessor, RnnToCnnPreProcessor, RnnToFeedForwardPreProcessor, TensorFlowCnnToFeedForwardPreProcessor, UnitVarianceProcessor, ZeroMeanAndUnitVariancePreProcessor, ZeroMeanPrePreProcessor

public interface InputPreProcessor
extends java.io.Serializable, java.lang.Cloneable

Input pre processor used for pre processing input before passing it to the neural network.

Method Summary

Modifier and Type	Method and Description
org.nd4j.linalg.api.ndarray.INDArray	backprop(org.nd4j.linalg.api.ndarray.INDArray output, int miniBatchSize) Reverse the preProcess during backprop.
InputPreProcessor	clone()
Pair<org.nd4j.linalg.api.ndarray.INDArray,MaskState>	feedForwardMaskArray(org.nd4j.linalg.api.ndarray.INDArray maskArray, MaskState currentMaskState, int minibatchSize)
InputType	getOutputType(InputType inputType) For a given type of input to this preprocessor, what is the type of the output?
org.nd4j.linalg.api.ndarray.INDArray	preProcess(org.nd4j.linalg.api.ndarray.INDArray input, int miniBatchSize) Pre preProcess input/activations for a multi layer network

Method Detail

preProcess

org.nd4j.linalg.api.ndarray.INDArray preProcess(org.nd4j.linalg.api.ndarray.INDArray input,
                                                int miniBatchSize)

Pre preProcess input/activations for a multi layer network

Parameters:

input - the input to pre preProcess

miniBatchSize -

Returns:

the processed input

backprop

org.nd4j.linalg.api.ndarray.INDArray backprop(org.nd4j.linalg.api.ndarray.INDArray output,
                                              int miniBatchSize)

Reverse the preProcess during backprop. Process Gradient/epsilons before passing them to the layer below.

Parameters:

output - which is a pair of the gradient and epsilon

miniBatchSize -

Returns:

the reverse of the pre preProcess step (if any)

clone

InputPreProcessor clone()

getOutputType

InputType getOutputType(InputType inputType)

For a given type of input to this preprocessor, what is the type of the output?

Parameters:

inputType - Type of input for the preprocessor

Returns:

Type of input after applying the preprocessor

feedForwardMaskArray

Pair<org.nd4j.linalg.api.ndarray.INDArray,MaskState> feedForwardMaskArray(org.nd4j.linalg.api.ndarray.INDArray maskArray,
                                                                          MaskState currentMaskState,
                                                                          int minibatchSize)