D3D：解耦扩散草稿异步推测解码技术

D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding

要点速览

D3D 通过独立的扩散草稿器打破对目标模型隐藏状态的依赖，实现异步草稿与验证。在七个基准测试中，D3D 结合块状去噪和辅助自回归监督，使 Qwen3-8B 和 Qwen3-32B 的平均解码速度分别提升 5.00 倍和 5.71 倍，优于 DFlash（4.03 倍和 3.67 倍），并在 GSM8K 上达到 7.48 倍加速。代码、模型检查点和基准测试脚本已开源。

正文

🚀 Introducing D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding!

Existing diffusion-based speculative drafters rely on target-model hidden states, forcing drafting and verification to execute sequentially.

D3D breaks this dependency with a standalone diffusion drafter, enabling asynchronous drafting and verification through packed speculative decoding.

Key highlights:

Asynchronous speculation: Overlaps diffusion drafting with target verification, eliminating inter-round dependencies.
Effective drafting: Combines blockwise denoising with auxiliary autoregressive supervision to achieve long accepted prefixes.
Significant speedups: Achieves 5.00× and 5.71× average decoding speedups with Qwen3-8B and Qwen3-32B across seven benchmarks, outperforming DFlash (4.03× and 3.67×). Up to 7.48× on GSM8K!

📄 Paper: https://arxiv.org/abs/2611.06425

💻 Code: https://github.com/fudan-generative-vision/D3D

🤗 Models: https://huggingface.co/collections/fudan-generative-ai/d3d

We've released our code, model checkpoints, and benchmarking scripts.

We welcome feedback, questions, and discussions!