Did Polymathic AI and Princeton researchers release FlashMap for faster LLM speculative decoding?
UnconfirmedLast checked
Polymathic AIFirst seen
Verdict
MarkTechPost covers the FlashMap release; primary paper/repo not re-opened in this run.
Evidence timeline · 1 source · newest first
Unconfirmed
MarkTechPost reports that Polymathic AI, Princeton, and USC ISI researchers released FlashMap, a speculative-decoding drafter that proposes 16 tokens jointly in one pass and claims about 6.14× faster LLM decoding versus prior baselines. Primary paper/code pages were not re-opened in this run.