We combine recent advancements in end-to-end speech recognition to non-autoregressive automatic speech recognition. We push the limits of non-autoregressive state-of-the-art results for multiple datasets: LibriSpeech, Fisher+Switchboard and Wall Street Journal. Key to our recipe, we leverage CTC on giant Conformer neural network architectures with SpecAugment and wav2vec2 pre-training. We achieve 1.8%/3.6% WER on LibriSpeech test/test-other sets, 5.1%/9.8% WER on Switchboard, and 3.4% on the Wall Street Journal, all without a language model.
@article{arxiv.2104.03416,
title = {Pushing the Limits of Non-Autoregressive Speech Recognition},
author = {Edwin G. Ng and Chung-Cheng Chiu and Yu Zhang and William Chan},
journal= {arXiv preprint arXiv:2104.03416},
year = {2021}
}