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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Self-supervised speech representation learning (SSL) has shown to beeffective in various downstream tasks, but SSL models are usually large andslow. Model compression techniques such as pruning aim to reduce the model sizeand computation without degradation in accuracy. Prior studies focus on thepruning of Transformers; however, speech models not only utilize a stack ofTransformer blocks, but also combine a frontend network based on multipleconvolutional layers for low-level feature representation learning. Thisfrontend has a small size but a heavy computational cost. In this work, wepropose three task-specific structured pruning methods to deal with suchheterogeneous networks. Experiments on LibriSpeech and SLURP show that theproposed method is more accurate than the original wav2vec2-base with 10% to30% less computation, and is able to reduce the computation by 40% to 50%without any degradation.

Structured Pruning of Self-Supervised Pre-trained Models for Speech Recognition and Understanding