---
title: "ChebBooster：基于切比雪夫外推的高效扩散Transformer免训练推理加速方法"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.23429v1"
permalink: "https://feed.astralor.com/items/itm_8593690f937a61844df53d2e"
published: "2026-08-24T16:06:11.000Z"
score: 75
channel: "ai"
tags: ["ChebBooster","切比雪夫多项式","图像生成","扩散Transformer","无训练加速"]
---

# ChebBooster：基于切比雪夫外推的高效扩散Transformer免训练推理加速方法

## AI 点评

> 我们认为ChebBooster的亮点不在于单纯的外推技巧，而是把数值稳定性问题变成了工程设计问题；离线权重预计算让它在落地时几乎不增加在线开销。对于跑DiT推理的工程师，这是一个可以直接尝试的加速选项，也值得观察它与蒸馏、量化等方法的叠加效果。

## 速览

扩散Transformer（DiT）虽图像生成质量高，但推理开销大。现有缓存加速长间隔精度低，泰勒外推则易受Runge振荡干扰。ChebBooster基于切比雪夫多项式提出无训练外推框架，用重心插值保证数值稳定、开销极小，并将过程拆为离线权重预计算与在线轻量应用。在DiT-XL/2、PixArt-Σ、FLUX.1-dev上最高实现3.68倍加速与5.12倍FLOPs降低，优于现有无训练基线。

## 关键点

- 提出基于切比雪夫多项式外推的无训练加速框架，避免训练成本
- 采用重心插值公式，解决数值不稳定性并保持极低在线开销
- 在DiT-XL/2、PixArt-Σ和FLUX.1-dev上最高实现3.68倍延迟加速和5.12倍FLOPs减少
- 外推过程拆分为离线权重预计算和在线轻量应用两阶段，便于实际部署

## 全文

Diffusion Transformers (DiTs) have shown strong performance in high-fidelity image generation, but their sampling process remains computationally intensive due to full model execution at every timestep. While cache-based acceleration has been explored to mitigate inference cost, naive reuse schemes suffer from low accuracy over long intervals, and Taylor-series-based extrapolation methods often face instability caused by Runge oscillations. In this paper, we propose ChebBooster, a training-free extrapolation framework based on Chebyshev polynomial theory that achieves stable and efficient acceleration for DiTs. Specifically, we adopt the Barycentric formulation to evaluate Chebyshev approximants with high numerical stability and minimal overhead, and further decouple the extrapolation into an offline weight precomputation phase and a lightweight online application stage. Extensive experiments across three representative DiT-based models, including DiT-XL/2, PixArt-$Σ$, and FLUX.1-dev, demonstrate that ChebBooster achieves consistent improvements in visual quality and inference efficiency, reaching up to $3.68\times$ latency speedup and $5.12\times$ FLOPs reduction, outperforming existing training-free baselines under diverse generation tasks and resolutions.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_8593690f937a61844df53d2e