API Reference - EncodingLayers

Constructors

OneHotEncoding

EncodingLayers.OneHotEncoding{tensor: tensor, finalDimensionSize: number, oneHotEncodingMode: string, indexDictionary: {any}}: AutomaticDifferentiationTensor

Parameters:

• tensor: The tensor that will be padded.

• finalDimensionSize: The final dimension size for the transformed tensor. It is equivalent to the number of labels that are available in the data.

• oneHotEncodingMode: The encoding mode to be used by the one hot encoding block. Available options are:

  • Index (Default)

  • Key

• indexDictionary: The index dictionary to be used to convert keys stored in the tensor to one hot encoding tensor. Must be given if using the “Key” one hot encoding mode.

Returns:

• AutomaticDifferentiationTensor: The automatic differentiation tensor that is created as a result of calling this function.

LabelEncoding

EncodingLayers.LabelEncoding{tensor: tensor, valueDictionary: {any}}: AutomaticDifferentiationTensor

Parameters:

• tensor: The tensor that will be padded.

• valueDictionary: The value dictionary to be used to convert keys stored in the tensor to values for label encoding tensor.

Returns:

• AutomaticDifferentiationTensor: The automatic differentiation tensor that is created as a result of calling this function.

PositionalEncoding

EncodingLayers.PositionalEncoding{tensor: tensor, sequenceLength: number, nValue: number}: AutomaticDifferentiationTensor

Parameters:

• tensor: The tensor that will be padded.

• sequenceLength: The length of the sequence.

• nValue: A user defined value for tuning.

Returns:

• AutomaticDifferentiationTensor: The automatic differentiation tensor that is created as a result of calling this function.