# Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducer Source: [Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducer]() Collection index: [llms.txt]() Data updated: 2025-10-25T20:52:10.607Z Authors: Kanishka Rao, Ha\\c{s}im Sak, Rohit Prabhavalkar Published: 2018-01-02T21:29:41.000Z ## Abstract We investigate training end-to-end speech recognition models with the recurrent neural network transducer (RNN-T): a streaming, all-neural, sequence-to-sequence architecture which jointly learns acoustic and language model components from transcribed acoustic data. We explore various model architectures and demonstrate how the model can be improved further if additional text or pronunciation data are available. The model consists of an \`encoder', which is initialized from a connectionist temporal classification-based (CTC) acoustic model, and a \`decoder' which is partially initialized from a recurrent neural network language model trained on text data alone. The entire neural network is trained with the RNN-T loss and directly outputs the recognized transcript as a sequence of graphemes, thus performing end-to-end speech recognition. We find that performance can be improved further through the use of sub-word units (\`wordpieces') which capture longer context and significantly reduce substitution errors. The best RNN-T system, a twelve-layer LSTM encoder with a two-layer LSTM decoder trained with 30,000 wordpieces as output targets achieves a word error rate of 8.5\\% on voice-search and 5.2\\% on voice-dictation tasks and is comparable to a state-of-the-art baseline at 8.3\\% on voice-search and 5.4\\% voice-dictation. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Speech Recognition on Turkish video dataset (clean)](): Speech Recognition; Turkish video dataset (clean); WER: 16.8. - [Speech Recognition on German video dataset (clean)](): Speech Recognition; German video dataset (clean); WER: 11.5. - [Speech Recognition on German video dataset (noisy)](): Speech Recognition; German video dataset (noisy); WER: 12.8. - [Speech Recognition on Turkish video dataset (noisy)](): Speech Recognition; Turkish video dataset (noisy); WER: 18.9. ## Original sources and implementation [Original paper]() [PDF]()