org.deeplearning4j.nn.conf

Class NeuralNetConfiguration.ListBuilder

java.lang.Object

org.deeplearning4j.nn.conf.MultiLayerConfiguration.Builder

org.deeplearning4j.nn.conf.NeuralNetConfiguration.ListBuilder

Enclosing class:

NeuralNetConfiguration

public static class NeuralNetConfiguration.ListBuilder
extends MultiLayerConfiguration.Builder

Fluent interface for building a list of configurations

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
class	NeuralNetConfiguration.ListBuilder.InputTypeBuilder Helper class for setting input types

Field Summary

Fields inherited from class org.deeplearning4j.nn.conf.MultiLayerConfiguration.Builder

backpropType, cacheMode, confs, dampingFactor, dataType, inferenceWorkspaceMode, inputPreProcessors, inputType, overrideNinUponBuild, tbpttBackLength, tbpttFwdLength, trainingWorkspaceMode, validateOutputConfig, validateTbpttConfig

Constructor Summary

Constructors

Constructor and Description
ListBuilder(NeuralNetConfiguration.Builder globalConfig)
ListBuilder(NeuralNetConfiguration.Builder globalConfig, Map<Integer,NeuralNetConfiguration.Builder> layerMap)

Method Summary

Modifier and Type	Method and Description
MultiLayerConfiguration.Builder	backpropType(@NonNull BackpropType type) The type of backprop.
MultiLayerConfiguration	build() Build the multi layer network based on this neural network and overr ridden parameters
NeuralNetConfiguration.ListBuilder	cacheMode(@NonNull CacheMode cacheMode) This method defines how/if preOutput cache is handled: NONE: cache disabled (default value) HOST: Host memory will be used DEVICE: GPU memory will be used (on CPU backends effect will be the same as for HOST)
NeuralNetConfiguration.ListBuilder	confs(List<NeuralNetConfiguration> confs)
NeuralNetConfiguration.ListBuilder	dataType(@NonNull DataType dataType) Set the DataType for the network parameters and activations for all layers in the network.
protected void	finalize()
List<InputType>	getLayerActivationTypes() For the (perhaps partially constructed) network configuration, return a list of activation sizes for each layer in the network. Note: To use this method, the network input type must have been set using setInputType(InputType) first
Map<Integer,NeuralNetConfiguration.Builder>	getLayerwise()
NeuralNetConfiguration.ListBuilder	inputPreProcessor(Integer layer, InputPreProcessor processor) Specify the processors.
NeuralNetConfiguration.ListBuilder	inputPreProcessors(Map<Integer,InputPreProcessor> processors)
NeuralNetConfiguration.ListBuilder.InputTypeBuilder	inputType() A convenience method for setting input types: note that for example .inputType().convolutional(h,w,d) is equivalent to .setInputType(InputType.convolutional(h,w,d))
NeuralNetConfiguration.ListBuilder	layer(int ind, @NonNull Layer layer)
NeuralNetConfiguration.ListBuilder	layer(Layer layer)
NeuralNetConfiguration.ListBuilder	overrideNinUponBuild(boolean overrideNinUponBuild) Whether to over ride the nIn configuration forcibly upon construction.
NeuralNetConfiguration.ListBuilder	setInputType(InputType inputType)
NeuralNetConfiguration.ListBuilder	tBPTTBackwardLength(int backwardLength) When doing truncated BPTT: how many steps of backward should we do? Only applicable when doing backpropType(BackpropType.TruncatedBPTT) This is the k2 parameter on pg23 of http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf
NeuralNetConfiguration.ListBuilder	tBPTTForwardLength(int forwardLength) When doing truncated BPTT: how many steps of forward pass should we do before doing (truncated) backprop? Only applicable when doing backpropType(BackpropType.TruncatedBPTT) Typically tBPTTForwardLength parameter is same as the tBPTTBackwardLength parameter, but may be larger than it in some circumstances (but never smaller) Ideally your training data time series length should be divisible by this This is the k1 parameter on pg23 of http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf
NeuralNetConfiguration.ListBuilder	tBPTTLength(int bpttLength) When doing truncated BPTT: how many steps should we do? Only applicable when doing backpropType(BackpropType.TruncatedBPTT) See: http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf
NeuralNetConfiguration.ListBuilder	validateOutputLayerConfig(boolean validate) Enabled by default.
NeuralNetConfiguration.ListBuilder	validateTbpttConfig(boolean validate) Enabled by default.

Methods inherited from class org.deeplearning4j.nn.conf.MultiLayerConfiguration.Builder

inferenceWorkspaceMode, trainingWorkspaceMode

Methods inherited from class java.lang.Object

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

Constructor Detail

ListBuilder

public ListBuilder(NeuralNetConfiguration.Builder globalConfig,
                   Map<Integer,NeuralNetConfiguration.Builder> layerMap)

ListBuilder

public ListBuilder(NeuralNetConfiguration.Builder globalConfig)

Method Detail

layer

public NeuralNetConfiguration.ListBuilder layer(int ind,
                                                @NonNull
                                                @NonNull Layer layer)

layer

public NeuralNetConfiguration.ListBuilder layer(Layer layer)

getLayerwise

public Map<Integer,NeuralNetConfiguration.Builder> getLayerwise()

overrideNinUponBuild

public NeuralNetConfiguration.ListBuilder overrideNinUponBuild(boolean overrideNinUponBuild)

Description copied from class: MultiLayerConfiguration.Builder

Whether to over ride the nIn configuration forcibly upon construction. Default value is true

Overrides:

overrideNinUponBuild in class MultiLayerConfiguration.Builder

Parameters:

overrideNinUponBuild - Whether to over ride the nIn configuration forcibly upon construction.

Returns:

builder pattern

inputPreProcessor

public NeuralNetConfiguration.ListBuilder inputPreProcessor(Integer layer,
                                                            InputPreProcessor processor)

Description copied from class: MultiLayerConfiguration.Builder

Specify the processors. These are used at each layer for doing things like normalization and shaping of input.

Overrides:

inputPreProcessor in class MultiLayerConfiguration.Builder

processor - what to use to preProcess the data.

Returns:

builder pattern

inputPreProcessors

public NeuralNetConfiguration.ListBuilder inputPreProcessors(Map<Integer,InputPreProcessor> processors)

Overrides:

inputPreProcessors in class MultiLayerConfiguration.Builder

cacheMode

public NeuralNetConfiguration.ListBuilder cacheMode(@NonNull
                                                    @NonNull CacheMode cacheMode)

Description copied from class: MultiLayerConfiguration.Builder

This method defines how/if preOutput cache is handled: NONE: cache disabled (default value) HOST: Host memory will be used DEVICE: GPU memory will be used (on CPU backends effect will be the same as for HOST)

Overrides:

cacheMode in class MultiLayerConfiguration.Builder

Returns:

backpropType

public MultiLayerConfiguration.Builder backpropType(@NonNull
                                                    @NonNull BackpropType type)

Description copied from class: MultiLayerConfiguration.Builder

The type of backprop. Default setting is used for most networks (MLP, CNN etc), but optionally truncated BPTT can be used for training recurrent neural networks. If using TruncatedBPTT make sure you set both tBPTTForwardLength() and tBPTTBackwardLength()

Overrides:

backpropType in class MultiLayerConfiguration.Builder

tBPTTLength

public NeuralNetConfiguration.ListBuilder tBPTTLength(int bpttLength)

Description copied from class: MultiLayerConfiguration.Builder

When doing truncated BPTT: how many steps should we do?
Only applicable when doing backpropType(BackpropType.TruncatedBPTT)
See: http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf

Overrides:

tBPTTLength in class MultiLayerConfiguration.Builder

Parameters:

bpttLength - length > 0

tBPTTForwardLength

public NeuralNetConfiguration.ListBuilder tBPTTForwardLength(int forwardLength)

Description copied from class: MultiLayerConfiguration.Builder

When doing truncated BPTT: how many steps of forward pass should we do before doing (truncated) backprop?
Only applicable when doing backpropType(BackpropType.TruncatedBPTT)
Typically tBPTTForwardLength parameter is same as the tBPTTBackwardLength parameter, but may be larger than it in some circumstances (but never smaller)
Ideally your training data time series length should be divisible by this This is the k1 parameter on pg23 of http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf

Overrides:

tBPTTForwardLength in class MultiLayerConfiguration.Builder

Parameters:

forwardLength - Forward length > 0, >= backwardLength

tBPTTBackwardLength

public NeuralNetConfiguration.ListBuilder tBPTTBackwardLength(int backwardLength)

Description copied from class: MultiLayerConfiguration.Builder

When doing truncated BPTT: how many steps of backward should we do?
Only applicable when doing backpropType(BackpropType.TruncatedBPTT)
This is the k2 parameter on pg23 of http://www.cs.utoronto.ca/~ilya/pubs/ilya_sutskever_phd_thesis.pdf

Overrides:

tBPTTBackwardLength in class MultiLayerConfiguration.Builder

Parameters:

backwardLength - <= forwardLength

confs

public NeuralNetConfiguration.ListBuilder confs(List<NeuralNetConfiguration> confs)

Overrides:

confs in class MultiLayerConfiguration.Builder

validateOutputLayerConfig

public NeuralNetConfiguration.ListBuilder validateOutputLayerConfig(boolean validate)

Description copied from class: MultiLayerConfiguration.Builder

Enabled by default. If enabled, the output layer configuration will be validated, to throw an exception on likely invalid outputs - such as softmax + nOut=1, or LossMCXENT + Tanh.
If disabled (false) no output layer validation will be performed.
Disabling this validation is not recommended, as the configurations that fail validation usually will not be able to learn correctly. However, the option to disable this validation is provided for advanced users when creating non-standard architectures.

Overrides:

validateOutputLayerConfig in class MultiLayerConfiguration.Builder

Parameters:

validate - If true: validate output layer configuration. False: don't validate

validateTbpttConfig

public NeuralNetConfiguration.ListBuilder validateTbpttConfig(boolean validate)

Description copied from class: MultiLayerConfiguration.Builder

Enabled by default. If enabled, an exception will be throw when using the (invalid) combination of truncated backpropagation through time (TBPTT) with either a GlobalPoolingLayer or LastTimeStepLayer.
It is possible to disable this validation to allow what is almost certainly an invalid configuration to be used, however this is not recommended.

Overrides:

validateTbpttConfig in class MultiLayerConfiguration.Builder

Parameters:

validate - Whether TBPTT validation should be performed

dataType

public NeuralNetConfiguration.ListBuilder dataType(@NonNull
                                                   @NonNull DataType dataType)

Description copied from class: MultiLayerConfiguration.Builder

Set the DataType for the network parameters and activations for all layers in the network. Default: Float

Overrides:

dataType in class MultiLayerConfiguration.Builder

Parameters:

dataType - Datatype to use for parameters and activations

finalize

protected void finalize()
                 throws Throwable

Overrides:

finalize in class Object

Throws:

Throwable

setInputType

public NeuralNetConfiguration.ListBuilder setInputType(InputType inputType)

Overrides:

setInputType in class MultiLayerConfiguration.Builder

inputType

public NeuralNetConfiguration.ListBuilder.InputTypeBuilder inputType()

A convenience method for setting input types: note that for example .inputType().convolutional(h,w,d) is equivalent to .setInputType(InputType.convolutional(h,w,d))

getLayerActivationTypes

public List<InputType> getLayerActivationTypes()

For the (perhaps partially constructed) network configuration, return a list of activation sizes for each layer in the network.
Note: To use this method, the network input type must have been set using setInputType(InputType) first

Returns:

A list of activation types for the network, indexed by layer number

build

public MultiLayerConfiguration build()

Build the multi layer network based on this neural network and overr ridden parameters

Overrides:

build in class MultiLayerConfiguration.Builder

Returns:

the configuration to build