Skip to content
AI News HubLIVE
More
Summary only

Polymathic AI and Princeton Researchers Release FlashMap, a Flow Map Drafter for 6.14x Faster LLM Decoding

Researchers from Polymathic AI, Princeton University, and USC ISI have released FlashMap, a speculative decoding drafter that replaces discrete diffusion with a continuous flow map. It drafts 16 tokens jointly in one forward pass and reaches up to 6.14x lossless speedup on Qwen3-8B.

SourceMarkTechPostAuthor: Asif Razzaq
Polymathic AI and Princeton Researchers Release FlashMap, a Flow Map Drafter for 6.14x Faster LLM Decoding
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions

Summary only

Under the source content permission, this page provides an editorial summary, key points, and a link to the original article.

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • FlashMap uses a 0.5B-parameter drafter that drafts 16-token blocks jointly in one pass, reaching up to 6.14x lossless speedup on Qwen3-8B
  • One-step FlashMap is faster than a retrained DFlash on all 9 Qwen3-8B benchmarks with roughly half the parameters
  • Weights are released under Apache-2.0 with MIT-licensed code and vLLM and SGLang plugins

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models