RNNs-implementation

Implementations of various Recurrent Neural Networks in Tensorflow 2

Content

• basic_builder.py: Returns specified RNN neural network
• custom_layers.py: Implementation of dilated RNN layer
• predefined_networks.py: Examples of networks specifications based on M4 competition winning solution by S. Smyl
• rnn_cells.py: Implementations of various RNN cells
• rnn_constructor.py: Given the RNN network specification given by either basic_builder or one of the predefined_networks this class constructs tf.keras.Model which can then be compiled and trained
• train_predefined_networks.ipynb: Simple example of how this repository can be used