Kangaroo introduces a self-speculative decoding framework for accelerating large language model inference, using a shallow sub-network and early exiting mechanisms to improve efficiency.

https://arxiv.org/abs//2404.18911

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