API Reference - Containers - RecurrentNeuralNetwork (RNN)

Constructors

new()

Creates a new cost function object.


RecurrentNeuralNetwork.new({inputDimensionSize: number, hiddenDimensionSize: number, learningRate: number, reverse: boolean}): ContainerObject

Parameters:

  • inputDimensionSize: The number of input features that will be accepted by the recurrent neural network cell.

  • hiddenDimensionSize: The number of hidden features that will be generated by the recurrent neural network cell.

  • learningRate: The learning rate that will be used by all the weight blocks that are stored in this recurrent neural network cell.

  • reverse: Determines whether the predictions should be generated starting from the end of the sequence. [Default: False]

Returns:

  • Container: The generated container object.

Functions

clearAllStoredTensors() 


RecurrentNeuralNetwork:clearAllStoredTensors()

Inherited From