ND
NVIDIA Developer
· 1 min read
VideosNVIDIA Developer (YouTube)
What is speculative decoding?
Maor Ashkenazi, research team lead at NVIDIA, explains how speculative decoding speeds up language model inference by drafting tokens ahead and having the full model verify or correct them.
Original source
This video was published by NVIDIA Developer (YouTube) and written by NVIDIA Developer. SyncAI.news embeds the publisher's own player.
Watch on youtube.com
