D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding

Target-conditioned diffusion drafters predict candidate tokens in parallel, yet each new block must wait for updated target-model hidden states.

By Chen, Mei, Lin +4Nov 4, 2026

Score█████░░░░░4.8

Key numbers

5.00× and 5.71
4.03× and 3.67
99% of mean acceptance length

Caveats

Preprint; not yet peer reviewed.

VerdictWorth a reader's time today.

Read the original PDF Code

Abstract

Target-conditioned diffusion drafters predict candidate tokens in parallel, yet each new block must wait for updated target-model hidden states. This forces drafting and verification to run sequentially, limiting speedup even though each block is generated in one pass. We introduce D3D, a standalone diffusion drafter that removes this inter-round dependency. D3D combines blockwise denoising with auxiliary autoregressive supervision to achieve long accepted prefixes without target hidden states. This independence lets D3D draft successor blocks during verification in one packed pass, enabling asynchronous speculative decoding. We progressively prune layers according to their effect on acceptance length to reduce the cost of packed drafting. Removing five of 36 layers retains over 99% of mean acceptance length across three benchmarks with Qwen3-8B as the target. Across seven benchmarks, D3D averages 5.00× and 5.71× decoding speedup with Qwen3-8B and Qwen3-32B targets, compared with 4.03× and 3.67× for DFlash. On GSM8K with Qwen3-32B, speedup reaches 7.48×. Code is publicly available at https://github.com/fudan-generative-vision/D3D.

Baoyou Chen, Ruiqiao Mei, Guancong Lin, Tao Song, Weihao Yuan, Qun Liu, Siyu Zhu

The editor's rubric

Heuristic review

Leverage4 / 5: A general-purpose tool used across several fields (Adam, ResNet, LoRA, next-generation sequencing).; 24% of this desk’s rubric
Magnitude3 / 5: Large gain: roughly 2x, or a clear new state of the art on a hard, unsaturated problem.; 18% of this desk’s rubric
Evidence3 / 5: Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data.; 14% of this desk’s rubric
Novelty2 / 5: A new combination of known ideas.; 16% of this desk’s rubric
Trajectory3 / 5: A clear path to scale.; 18% of this desk’s rubric
Stakes2 / 5: Benefits a professional community (practitioners, clinicians, engineers).; 10% of this desk’s rubric

Dimension	Level	Weight	What that level means
Leverage	████░4/5	24%	A general-purpose tool used across several fields (Adam, ResNet, LoRA, next-generation sequencing).
Magnitude	███░░3/5	18%	Large gain: roughly 2x, or a clear new state of the art on a hard, unsaturated problem.
Evidence	███░░3/5	14%	Solid: multiple benchmarks or cohorts, ablations, fair baselines, released code or data.
Novelty	██░░░2/5	16%	A new combination of known ideas.
Trajectory	███░░3/5	18%	A clear path to scale.
Stakes	██░░░2/5	10%	Benefits a professional community (practitioners, clinicians, engineers).

Editor’s rationale

Heuristic triage from title and abstract text only, not a reading of the paper. Cues found: method (we propose); breadth (many tasks); gains (x-fold); verification (multiple benchmarks, code released); scale (efficient).

How the score was computed

rank-2026-10-18

Score█████░░░░░4.8

Score = 10 × (65% × adjusted merit / 10 + 25% × attention + 10% × freshness)

Desks place stronger verdicts first, then use this score to order work within the same verdict.

Merit: 6.0 / 10; Weighted rubric, evidence-gated.
Adjusted merit: 4.8 / 10; Shrunk toward the desk prior by editor confidence (40%).
Attention: 47%; Citations, upvotes, points, mentions.
Freshness: 57%; Half-life decay since publication.

Hugging Face upvotes17 (reference 25, via hf-daily, Nov 6, 2026, 13:40 UTC)
GitHub stars19 (reference 250, via hf-daily, Nov 6, 2026, 13:40 UTC)

The record

Reviewed by heuristic-v10 on Nov 5, 2026, 13:49 UTC. Paper type: method.
Categories: cs.AI, cs.LG
BRIEF, No.1 in the Artificial Intelligence edition of November 6, 2026.