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

Recent progress in Automatic Speech Recognition (ASR) has been coupled with asubstantial increase in the model sizes, which may now contain billions ofparameters, leading to slow inferences even with adapted hardware. In thiscontext, several ASR models exist in various sizes, with different inferencecosts leading to different performance levels. Based on the observation thatsmaller models perform optimally on large parts of testing corpora, we proposeto train a decision module, that would allow, given an audio sample, to use thesmallest sufficient model leading to a good transcription. We apply ourapproach to two Whisper models with different sizes. By keeping the decisionprocess computationally efficient, we build a decision module that allowssubstantial computational savings with reduced performance drops.

Big model only for hard audios: Sample dependent Whisper model selection for efficient inferences