The paper presents Distil-Whisper, a smaller variant of the Whisper speech recognition model, achieved through distillation and pseudo-labelling. Distil-Whisper is faster and has fewer parameters while maintaining performance and robustness. Code and models are publicly available.

https://arxiv.org/abs//2311.00430

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers