from Hacker News

Implementing LLM Speculative Sampling in Under 100 Lines of Code

by mathewshen on 3/13/25, 6:46 AM with 0 comments