FlashMap Uses Continuous Flow Maps for Single-Pass Speculative Decoding
FlashMap is a new speculative decoding drafter that replaces discrete diffusion with continuous flow maps to jointly generate token blocks in a single forward pass. It achieves up to 6.14x lossless speedup on Qwen3-8B, outperforming retrained DFlash by ~7% on average with half the parameters (0.5B vs 1B). Apache-2.0 weights and MIT code are now available with vLLM and SGLang support.
Full Report
▲ Marktechpost AI · 2026-10-28 08
FlashMap, a speculative decoding drafter that writes a whole block of tokens jointly, in 1 forward pass.
The key idea is to swap discrete diffusion for a continuous flow map. Diffusion drafters like DFlash predict each slot independently, so drafts can mix "new" with "diego". FlashMap maps 1 noise sample to 1 full block, so the 16 tokens stay consistent.
On Qwen3-8B, 1-step FlashMap reaches up to 6.14x lossless speedup and beats a retrained DFlash on all 9 benchmarks, about 7% faster on average. It does this with a 0.5B drafter vs DFlash's 1B. It is a trade-off though: 1-step drafts lose acceptance at late block slots, and on Qwen3-4B it trails DFlash on 4 of 9 tasks at temperature 0.
Drafters for Qwen3-8B and Qwen3-4B are out now, Apache-2.0 weights and MIT code, with vLLM 0.31 and SGLang 0.5.20 plugins......
Full analysis:
