Polymathic AI and Princeton Researchers Release FlashMap, a Flow Map Drafter for 6.14x Faster LLM Decoding
Researchers from Polymathic AI, Princeton University, USC ISI, NYU and the Flatiron Institute have released FlashMap. FlashMap speculative decoding replaces the discrete diffusion…
Researchers from Polymathic AI, Princeton University, USC ISI, NYU and the Flatiron Institute have released FlashMap.FlashMap speculative decoding replaces the discrete diffusion drafter with a continuous flow map.
It drafts a full 16-token block in 1 forward pass.Tokens in that block stay consistent with each other.On Qwen3-8B, the research team reports up to 6.14x lossless speedup on GSM8K.Drafter weights for Qwen3-8B and Qwen3-4B are on Hugging Face, with vLLM and SGLang plugins.
FlashMap is a speculative decoding drafter that maps Gaussian noise to a token block in 1 step. A small drafter proposes K tokens. The target model then checks all K in 1 parallel pass and keeps the longest matching prefix. Output stays exactly what the target would produce.
Speed depends on 2 things. How many drafted tokens get accepted, and how cheaply they were drafted. EAGLE-3 drafts autoregressively, so cost grows with block length. DFlash drafts the full block in 1 pass with block diffusion.
DFlash predicts each slot in the block independently. The research team calls this the factorized approximation. Picture a continuation that is “new york” or “san diego” with equal odds. An independent drafter proposes “new diego” or “san york” half the time. The target rejects those at slot 2.
FlashMap builds on Flow Map Language Models. Each noise sample maps to 1 full block. The noise picks the mode for every slot jointly.
The drafter is a 6-block diffusion transformer (DiT), with embedding dimension 2048 and 16 attention heads. It reads hidden states from target layers 1, 9, 17, 25 and 33. Embedding matrix and LM head are copied from the target and stay frozen.
Both drafters train on 500K samples from Open-PerfectBlend. Total compute across all models and ablations was about 4,600 GPU hours on H200s.
The research team retrained EAGLE-3 and DFlash on the same data and settings. Tests cover 9 benchmarks across math, code and chat, at temperatures 0 and 1.
A shuffle test checks whether drafts use real inter-token structure. At 2 steps, FlashMap recovers most of the joint-structure gain of a 16-step teacher. At 1 step, it keeps only a small fraction. Drafting 4 noise samples as a tree raised speedup by 11% on average.
Average speedups are computed by Marktechpost from Figure 1(c) of the FlashMap paper. Matched rows use the authors’ retrained baselines.
The flashmap package needs only torch and transformers>=4.56. Install it with uv pip install flashmap. A spec_generate call runs drafter and target together on 1 GPU.
For serving, the project page gives 2 paths:
Set num_steps=2 to trade some speed for higher acceptance. For sampling above temperature 0, vLLM needs "draft_sample_method": "probabilistic".
Check out the paper, project page, GitHub repo and model weights. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us
Asif Razzaq is the CEO of Marktechpost AI Media Inc..As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good.His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience.
The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.
Supercharge your agentic search with Decision‑powered retrieval
The newsletter 70,000+ AI builders read. Model releases, tools and research, without the hype.
Practitioner-first AI/ML news and analysis, read by 1M+ developers and researchers every month.