Normalization Layers
KerasBatchNormalization
Imports a BatchNormalization layer from Keras.
KerasBatchNormalization
public KerasBatchNormalization(Integer kerasVersion) throws UnsupportedKerasConfigurationException
Pass-through constructor from KerasLayer
param kerasVersion major keras version
throws UnsupportedKerasConfigurationException Unsupported Keras config
getBatchNormalizationLayer
public BatchNormalization getBatchNormalizationLayer()
Constructor from parsed Keras layer configuration dictionary.
param layerConfig dictionary containing Keras layer configuration
throws InvalidKerasConfigurationException Invalid Keras config
throws UnsupportedKerasConfigurationException Unsupported Keras config
getOutputType
public InputType getOutputType(InputType... inputType) throws InvalidKerasConfigurationException
Get layer output type.
param inputType Array of InputTypes
return output type as InputType
throws InvalidKerasConfigurationException Invalid Keras config
getNumParams
public int getNumParams()
Returns number of trainable parameters in layer.
return number of trainable parameters (4)
setWeights
public void setWeights(Map<String, INDArray> weights) throws InvalidKerasConfigurationException
Set weights for layer.
param weights Map from parameter name to INDArray.
Last updated
Was this helpful?