English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The intensive computation of Automatic Speech Recognition (ASR) modelsobstructs them from being deployed on mobile devices. In this paper, we presenta novel quantized Winograd optimization pipeline, which combines thequantization and fast convolution to achieve efficient inference accelerationon mobile devices for ASR models. To avoid the information loss due to thecombination of quantization and Winograd convolution, a Range-ScaledQuantization (RSQ) training method is proposed to expand the quantizednumerical range and to distill knowledge from high-precision values. Moreover,an improved Conv1D equipped DFSMN (ConvDFSMN) model is designed for mobiledeployment. We conduct extensive experiments on both ConvDFSMN and Wav2lettermodels. Results demonstrate the models can be effectively optimized with theproposed pipeline. Especially, Wav2letter achieves 1.48* speedup with anapproximate 0.07% WER decrease on ARMv7-based mobile devices.

INT8 Winograd Acceleration for Conv1D Equipped ASR Models Deployed on Mobile Devices