RETURNN: The RWTH Extensible Training framework for Universal Recurrent Neural Networks
This provides a configurable tool for researchers to train advanced RNN models, but it is incremental as it builds on existing software frameworks.
The authors released RETURNN, an extensible training framework for recurrent neural networks, which enables efficient multi-GPU training of state-of-the-art deep bidirectional LSTM models on data like speech and handwritten text, used in successful evaluation campaigns.
In this work we release our extensible and easily configurable neural network training software. It provides a rich set of functional layers with a particular focus on efficient training of recurrent neural network topologies on multiple GPUs. The source of the software package is public and freely available for academic research purposes and can be used as a framework or as a standalone tool which supports a flexible configuration. The software allows to train state-of-the-art deep bidirectional long short-term memory (LSTM) models on both one dimensional data like speech or two dimensional data like handwritten text and was used to develop successful submission systems in several evaluation campaigns.