Returnn: The RWTH extensible training framework for universal recurrent neural networks

By Patrick Doetsch, Albert Zeyer, Paul Voigtlaender, Ilya Kulikov, Ralf Schlüter, Hermann Ney. 2016

79 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing, Model Architectures

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Natural Language Processing

In Model Architectures

Citations in the corpus

RWTH ASR Systems for LibriSpeech: Hybrid vs Attention - w/o Data Augmentation (2019, methodology): “The hybrid model was trained and decoded with RASR [38] and RETURNN [23, 39].”
Improved training of end-to-end attention models for speech recognition (2018, methodology): “We use RETURNN [30,45] based on TensorFlow [46] for its computation.”
Generating Synthetic Audio Data for Attention-Based Speech Recognition Systems (2019, methodology): “The experiments in this work were managed with Sisyphus [19] and the ASR and TTS systems were implemented in RETURNN [20].”
Bidirectional Decoder Networks for Attention-Based End-to-End Offline Handwriting Recognition (2016, methodology): “We used RETURNN [26], the recurrent neural network training software developed at our institute, to train all networks and an example setup to run these experiments will be provided.”
Recent Advances in End-to-End Automatic Speech Recognition (2021, background): “First of all, great toolkits [56, 227, 252, 297, 299, 310] are indispensable to the fast development of E2E ASR”
End-to-End Speech Recognition: A Survey (2023, methodology): “The end-to-end ASR toolkit including ESPnet6[215], ESPRESSO7[218], and RETURNN8[228] also support the vectorized beam search algorithm.”
A comprehensive study of deep bidirectional LSTM RNNS for acoustic modeling in speech recognition (2016, methodology): “We try to fill this gap by a comprehensive study of various aspects of training deep BLSTMs and we provide configuration files for all our experiments [5] for our framework RETURNN [6].”
The Rwth Asr System for Ted-Lium Release 2: Improving Hybrid Hmm With Specaugment (2020, methodology): “All trainings are done using our NN modeling toolkit RETURNN [15], our ASR toolkit RASR [16] and our work-ﬂow manager Sisyphus [17].”
The RWTH/UPB/FORTH System Combination for the 4th CHiME Challenge Evaluation (2016, methodology): “The ﬁrst back end was implemented using RWTH’s open-source toolkits RASR [5] and RETURNN [6].”
RETURNN as a Generic Flexible Neural Toolkit with Application to Translation and Speech Recognition (2018, background): “RETURNN, the RWTH extensible training framework for universal recurrent neural networks, was introduced in (Doetsch et al., 2017).”