# Uno 用扩散并行草稿提速 LLM 推理

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-09-09 01:37
- AIHOT 分数：43
- AIHOT 链接：https://aihot.news/items/cmtsyggsz01r8ro5wxoo1l79n
- 原文链接：https://x.com/rohanpaul_ai/status/2097378878000157076

## AI 摘要

Uno 提出一种不改变输出分布的 LLM 加速方案：保留原自回归模型负责质量，仅用扩散模型并行草拟多个 token 供其验证，无需独立草稿模型。在 Qwen3-8B 上，最大测试批次下请求吞吐量提升 2.5 倍，系统吞吐量提升 1.6 倍，端到端 RL 训练提速最高 40%。

## 正文

Uno shows a simple way to speed up existing LLMs without changing their output distribution: keep the original model in charge, and use diffusion only to draft multiple tokens in parallel.

Keep the autoregressive model for quality and add diffusion for parallel drafting

On Qwen3-8B, Uno delivered 2.5X higher per-request throughput and 1.6X higher system throughput at the largest tested batch size.

Normal autoregressive LLMs generate 1 token at a time, making long responses, agents, and RL rollouts slow.

Uno adds lightweight diffusion adapters that propose several future tokens at once, while the original model verifies them.

That avoids a separate draft model and preserves the base model’s sampling behavior.

It also sped up end-to-end RL training by up to 40% in the reported runs.