mlsanityData to 7 Oct 2026

RETURNN as a Generic Flexible Neural Toolkit with Application to Translation and Speech Recognition

arXiv 1805.05225 ↗First posted 14 May 2018cs.NE cs.AI cs.CLcited by 30

Not on the map. Why?

Abstract (arXiv)

We compare the fast training and decoding speed of RETURNN of attention models for translation, due to fast CUDA LSTM kernels, and a fast pure TensorFlow beam search decoder. We show that a layer-wise pretraining scheme for recurrent attention models gives over 1% BLEU improvement absolute and it allows to train deeper recurrent encoder networks. Promising preliminary results on max. expected BLEU training are presented. We are able to train state-of-the-art models for translation and end-to-end models for speech recognition and show results on WMT 2017 and Switchboard. The flexibility of RETURNN allows a fast research feedback loop to experiment with alternative architectures, and its generality allows to use it on a wide range of applications.

Key facts. Cited by 30 papers in the map (data through 7 Oct 2026). First posted on arXiv 14 May 2018.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 8; 2022: 3; 2023: 10; 2024: 5; 2025: 4.

30 papers.

Sort:

  1. 2021SpeechBrain: A General-Purpose Speech Toolkitcited by 404
  2. 2023End-to-End Speech Recognition: A Surveycited by 127
  3. 2021Why does CTC result in peaky behavior?cited by 32
  4. 2021Librispeech Transducer Model with Internal Language Model Prior Correctioncited by 26
  5. 2021Investigating Methods to Improve Language Model Integration for Attention-based Encoder-Decoder ASR Modelscited by 25
  6. 2021Conformer-based Hybrid ASR System for Switchboard Datasetcited by 15
  7. 2022Improving the Training Recipe for a Robust Conformer-based Hybrid Modelcited by 9
  8. 2024VietMed: A Dataset and Benchmark for Automatic Speech Recognition of Vietnamese in the Medical Domaincited by 9
  9. 2023Chunked Attention-based Encoder-Decoder Model for Streaming Speech Recognitioncited by 7
  10. 2024MultiMed: Multilingual Medical Speech Recognition via Attention Encoder Decodercited by 7
  11. 2021A study of latent monotonic attention variantscited by 5
  12. 2022Monotonic segmental attention for automatic speech recognitioncited by 4
18 more
  1. 2023Mixture Encoder for Joint Speech Separation and Recognitioncited by 4
  2. 2023Comparative Analysis of the wav2vec 2.0 Feature Extractorcited by 4
  3. 2022AppTek's Submission to the IWSLT 2022 Isometric Spoken Language Translation Taskcited by 3
  4. 2023RASR2: The RWTH ASR Toolkit for Generic Sequence-to-sequence Speech Recognitioncited by 3
  5. 2023Take the Hint: Improving Arabic Diacritization with Partially-Diacritized Textcited by 2
  6. 2024The Conformer Encoder May Reverse the Time Dimensioncited by 2
  7. 2025Supplementary Resources and Analysis for Automatic Speech Recognition Systems Trained on the Loquacious Datasetcited by 2
  8. 2021On Sampling-Based Training Criteria for Neural Language Modelingcited by 1
  9. 2023Improving Frame-level Classifier for Word Timings with Non-peaky CTC in End-to-End Automatic Speech Recognitioncited by 1
  10. 2023Developing automatic verbatim transcripts for international multilingual meetings: an end-to-end solutioncited by 1
  11. 2023Weak Alignment Supervision from Hybrid Model Improves End-to-end ASRcited by 1
  12. 2024wav2graph: A Framework for Supervised Learning Knowledge Graph from Speechcited by 1
  13. 2025Efficient Supernet Training with Orthogonal Softmax for Scalable ASR Model Compressioncited by 1
  14. 2025Regularizing Learnable Feature Extraction for Automatic Speech Recognitioncited by 1
  15. 2025Dynamic Acoustic Model Architecture Optimization in Training for ASRcited by 1
  16. 2021Automatic Learning of Subword Dependent Model Scalescited by 0
  17. 2023Analyzing And Improving Neural Speaker Embeddings for ASRcited by 0
  18. 2024Alternating Weak Triphone/BPE Alignment Supervision from Hybrid Model Improves End-to-End ASRcited by 0