org.deeplearning4j.nn.conf.layers

Class Layer

java.lang.Object

org.deeplearning4j.nn.conf.layers.Layer

All Implemented Interfaces:

Serializable, Cloneable

Direct Known Subclasses:

AbstractSameDiffLayer, ActivationLayer, BaseLayer, BaseUpsamplingLayer, BaseWrapperLayer, Bidirectional, FrozenLayer, GlobalPoolingLayer, LocalResponseNormalization, MaskLayer, NoParamLayer, SpaceToBatchLayer, SpaceToDepthLayer, Subsampling3DLayer, SubsamplingLayer, Yolo2OutputLayer, ZeroPadding1DLayer, ZeroPadding3DLayer

public abstract class Layer
extends Object
implements Serializable, Cloneable

A neural network layer.

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Layer.Builder<T extends Layer.Builder<T>>

Field Summary

Fields

Modifier and Type	Field and Description
protected List<LayerConstraint>	constraints
protected IDropout	iDropout
protected String	layerName

Constructor Summary

Constructors

Constructor and Description
Layer(Layer.Builder builder)

Method Summary

Modifier and Type	Method and Description
Layer	clone()
abstract double	getL1ByParam(String paramName) Get the L1 coefficient for the given parameter.
abstract double	getL2ByParam(String paramName) Get the L2 coefficient for the given parameter.
abstract LayerMemoryReport	getMemoryReport(InputType inputType) This is a report of the estimated memory consumption for the given layer
abstract InputType	getOutputType(int layerIndex, InputType inputType) For a given type of input to this layer, what is the type of the output?
abstract InputPreProcessor	getPreProcessorForInputType(InputType inputType) For the given type of input to this layer, what preprocessor (if any) is required? Returns null if no preprocessor is required, otherwise returns an appropriate InputPreProcessor for this layer, such as a CnnToFeedForwardPreProcessor
org.nd4j.linalg.learning.config.IUpdater	getUpdaterByParam(String paramName) Get the updater for the given parameter.
protected void	initializeConstraints(Layer.Builder<?> builder) Initialize the weight constraints.
abstract ParamInitializer	initializer()
abstract Layer	instantiate(NeuralNetConfiguration conf, Collection<TrainingListener> trainingListeners, int layerIndex, org.nd4j.linalg.api.ndarray.INDArray layerParamsView, boolean initializeParams)
abstract boolean	isPretrainParam(String paramName) Is the specified parameter a layerwise pretraining only parameter? For example, visible bias params in an autoencoder (or, decoder params in a variational autoencoder) aren't used during supervised backprop. Layers (like DenseLayer, etc) with no pretrainable parameters will return false for all (valid) inputs.
void	resetLayerDefaultConfig() Reset the learning related configs of the layer to default.
abstract void	setNIn(InputType inputType, boolean override) Set the nIn value (number of inputs, or input channels for CNNs) based on the given input type

Methods inherited from class java.lang.Object

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

Field Detail

layerName

protected String layerName

iDropout

protected IDropout iDropout

constraints

protected List<LayerConstraint> constraints

Constructor Detail

Layer

public Layer(Layer.Builder builder)

Method Detail

initializeConstraints

protected void initializeConstraints(Layer.Builder<?> builder)

Initialize the weight constraints. Should be called last, in the outer-most constructor

resetLayerDefaultConfig

public void resetLayerDefaultConfig()

Reset the learning related configs of the layer to default. When instantiated with a global neural network configuration the parameters specified in the neural network configuration will be used. For internal use with the transfer learning API. Users should not have to call this method directly.

clone

public Layer clone()

Overrides:

clone in class Object

instantiate

public abstract Layer instantiate(NeuralNetConfiguration conf,
                                  Collection<TrainingListener> trainingListeners,
                                  int layerIndex,
                                  org.nd4j.linalg.api.ndarray.INDArray layerParamsView,
                                  boolean initializeParams)

initializer

public abstract ParamInitializer initializer()

Returns:

The parameter initializer for this model

getOutputType

public abstract InputType getOutputType(int layerIndex,
                                        InputType inputType)

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

Parameters:

layerIndex - Index of the layer

inputType - Type of input for the layer

Returns:

Type of output from the layer

Throws:

IllegalStateException - if input type is invalid for this layer

setNIn

public abstract void setNIn(InputType inputType,
                            boolean override)

Set the nIn value (number of inputs, or input channels for CNNs) based on the given input type

Parameters:

inputType - Input type for this layer

override - If false: only set the nIn value if it's not already set. If true: set it regardless of whether it's already set or not.

Throws:

IllegalStateException - if input type is invalid for this layer

getPreProcessorForInputType

public abstract InputPreProcessor getPreProcessorForInputType(InputType inputType)

For the given type of input to this layer, what preprocessor (if any) is required?
Returns null if no preprocessor is required, otherwise returns an appropriate InputPreProcessor for this layer, such as a CnnToFeedForwardPreProcessor

Parameters:

inputType - InputType to this layer

Returns:

Null if no preprocessor is required, otherwise the type of preprocessor necessary for this layer/input combination

Throws:

IllegalStateException - if input type is invalid for this layer

getL1ByParam

public abstract double getL1ByParam(String paramName)

Get the L1 coefficient for the given parameter. Different parameters may have different L1 values, even for a single .l1(x) configuration. For example, biases generally aren't L1 regularized, even if weights are

Parameters:

paramName - Parameter name

Returns:

L1 value for that parameter

getL2ByParam

public abstract double getL2ByParam(String paramName)

Get the L2 coefficient for the given parameter. Different parameters may have different L2 values, even for a single .l2(x) configuration. For example, biases generally aren't L1 regularized, even if weights are

Parameters:

paramName - Parameter name

Returns:

L2 value for that parameter

isPretrainParam

public abstract boolean isPretrainParam(String paramName)

Is the specified parameter a layerwise pretraining only parameter?
For example, visible bias params in an autoencoder (or, decoder params in a variational autoencoder) aren't used during supervised backprop.
Layers (like DenseLayer, etc) with no pretrainable parameters will return false for all (valid) inputs.

Parameters:

paramName - Parameter name/key

Returns:

True if the parameter is for layerwise pretraining only, false otherwise

getUpdaterByParam

public org.nd4j.linalg.learning.config.IUpdater getUpdaterByParam(String paramName)

Get the updater for the given parameter. Typically the same updater will be used for all updaters, but this is not necessarily the case

Parameters:

paramName - Parameter name

Returns:

IUpdater for the parameter

getMemoryReport

public abstract LayerMemoryReport getMemoryReport(InputType inputType)

This is a report of the estimated memory consumption for the given layer

Parameters:

inputType - Input type to the layer. Memory consumption is often a function of the input type

Returns:

Memory report for the layer