Beta Version 1.6.0

Added

• Added RecurrentNeuralNetworkCell, RecurrentNeuralNetwork, GatedRecurrentUnitCell and GatedRecurrent Unit to the “Containers” section.

• Added Diffusion to the “Models” section.

• Added PairwiseDistance to the “OperatorBlocks” section.

Changes

• Some containers now have the ability to update the weights immediately after weight loss tensor calculation and is activated by default. This includes Sequential, RecurrentNeuralNetworkCell and GatedRecurrentUnitCell containers. This allows the containers to train more quickly compared to the standard implementation of backward propagation.

• Containers now inherits from BaseContainer class to reduce the maintenance effort.