org.deeplearning4j.nn.conf.layers.samediff

Class SDLayerParams

java.lang.Object

org.deeplearning4j.nn.conf.layers.samediff.SDLayerParams

All Implemented Interfaces:

Serializable

public class SDLayerParams
extends Object
implements Serializable

SDLayerParams is used to define the parameters for a Deeplearning4j SameDiff layer

Author:

Alex Black

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
SDLayerParams(Map<String,int[]> weightParams, Map<String,int[]> biasParams)

Method Summary

Modifier and Type	Method and Description
void	addBiasParam(String paramKey, int[] paramShape) Add a bias parameter to the layer, with the specified shape.
void	addWeightParam(String paramKey, int... paramShape) Add a weight parameter to the layer, with the specified shape.
void	clear() Clear any previously set weight/bias parameters (including their shapes)
boolean	equals(Object o)
List<String>	getBiasParameterKeys()
List<String>	getParameterKeys()
Map<String,int[]>	getParamShapes() Get the parameter shapes for all parameters
List<String>	getWeightParameterKeys()
int	hashCode()

Methods inherited from class java.lang.Object

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

Constructor Detail

SDLayerParams

public SDLayerParams(Map<String,int[]> weightParams,
                     Map<String,int[]> biasParams)

Method Detail

addWeightParam

public void addWeightParam(@NonNull
                           String paramKey,
                           @NonNull
                           int... paramShape)

Add a weight parameter to the layer, with the specified shape. For example, a standard fully connected layer could have weight parameters with shape [numInputs, layerSize]

Parameters:

paramKey - The parameter key (name) for the weight parameter

paramShape - Shape of the weight parameter array

addBiasParam

public void addBiasParam(@NonNull
                         String paramKey,
                         @NonNull
                         int[] paramShape)

Add a bias parameter to the layer, with the specified shape. For example, a standard fully connected layer could have bias parameters with shape [1, layerSize]

Parameters:

paramKey - The parameter key (name) for the bias parameter

paramShape - Shape of the bias parameter array

getParameterKeys

public List<String> getParameterKeys()

Returns:

Get a list of parameter names / keys (previously added via addWeightParam(String, int...) and addBiasParam(String, int[])

getWeightParameterKeys

public List<String> getWeightParameterKeys()

Returns:

Get a list of parameter names / keys for weight parameters only, previously added via addWeightParam(String, int...)

getBiasParameterKeys

public List<String> getBiasParameterKeys()

Returns:

Get a list of parameter names / keys for weight parameters only, previously added via addWeightParam(String, int...)

getParamShapes

public Map<String,int[]> getParamShapes()

Get the parameter shapes for all parameters

Returns:

Map of parameter shapes, by parameter

clear

public void clear()

Clear any previously set weight/bias parameters (including their shapes)

equals

public boolean equals(Object o)

Overrides:

equals in class Object

hashCode

public int hashCode()

Overrides:

hashCode in class Object