提出TCC方法，让缓存加速的扩散模型生成更准
Trajectory-Consistent Calibration for Cache-Accelerated Diffusion Models
- 用迭代校准法修复缓存带来的表示偏差
- 在PixArt-alpha上使FID从29.83降至27.35
- 无需训练，适配各类缓存加速方案
扩散Transformer在迭代采样中需反复评估去噪器，推理成本高。基于缓存的加速通过复用中间表示降低开销，但可能引入表示偏差，影响生成质量。本文分析此类偏差，指出有效校准需同时考虑直接重用失配和前期修正引发的轨迹偏移。为此提出训练无关的轨迹一致性校准（TCC）方法，将缓存表示校准至全计算版本。TCC采用离线迭代过程，使每个校准先验都考虑前序校正导致的轨迹变化。在PixArt-alpha和DiT-XL/2上的实验表明，TCC在多种代表性缓存加速方法中持续提升FID表现，且保留原有复用策略。尤其在基于FORA的PixArt-alpha缓存加速设置下，FID由29.83降至27.35，略优于全计算基线。
原文摘要 · Abstract (English)
Diffusion Transformers require repeated denoiser evaluations during iterative sampling, making inference computationally expensive. Cache-based acceleration reduces this cost by reusing intermediate representations across denoising steps, but can introduce representation deviations and degrade generation quality. In this paper, we analyze these deviations and show that effective calibration should consider both the direct mismatch caused by reuse and the subsequent trajectory shift induced by earlier corrections. To address this challenge, we propose Trajectory-Consistent Calibration (TCC), a training-free method that calibrates cached representations toward their full-computation counterparts. Specifically, rather than estimating all calibration priors from a single uncorrected cache trajectory, TCC uses an offline iterative procedure so that each prior accounts for the trajectory shift induced by preceding calibrations. Experiments on PixArt-alpha and DiT-XL/2 show that TCC consistently improves FID across representative cache-based acceleration methods while preserving their underlying reuse policies. Notably, in a representative PixArt-alpha cache-acceleration setting based on FORA, TCC reduces FID from 29.83 to 27.35, slightly surpassing the full-computation baseline.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。