Polymathic AI and Princeton Researchers Release FlashMap, a Flow Map Drafter for 6.14x Faster LLM Decoding
FlashMap speculative decoding uses a 0.5B flow map drafter to reach 6.14x lossless speedup on Qwen3-8B, beating DFlash.
MarkTechPost · Asif Razzaq · https://www.facebook.com/MarkTechPost/