通过感知敏感度动态缓存，提升扩散模型推理速度与质量。
SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching
- 基于输入扰动敏感度分析，动态选择缓存时间步。
- 在相同计算量下，生成视频质量优于现有缓存方法。
- 适用于需要高效推理的视频生成场景，如Cinematic Video Generation。
扩散模型在视频生成方面达到顶尖质量，但其推理成本高昂，主要源于大量顺序去噪步骤。为加速推理，研究者提出了多种无需训练的加速方法，其中缓存技术通过复用不同时间步的模型输出来减少计算量。然而，现有缓存方法依赖启发式规则选择缓存时间步，需大量调参。本文提出一种基于敏感度感知的缓存框架（SenCache），通过分析模型输出对噪声隐变量和时间步扰动的敏感度，证明该敏感度是缓存误差的关键预测因子。据此设计出动态缓存策略，可针对每个样本自适应选择缓存时间步。该框架为自适应缓存提供了理论基础，解释了先前经验性启发式方法的部分有效性，并将其推广为样本相关的动态方法。在Wan 2.1、CogVideoX和LTX-Video上的实验表明，相比现有方法，SenCache在相似计算开销下实现了更优的视觉质量。
原文摘要 · Abstract (English)
Diffusion models achieve state-of-the-art video generation quality, but their inference remains expensive due to the large number of sequential denoising steps. This has motivated a growing line of research on accelerating diffusion inference. Among training-free acceleration methods, caching reduces computation by reusing previously computed model outputs across timesteps. Existing caching methods rely on heuristic criteria to choose cache/reuse timesteps and require extensive tuning. We address this limitation with a principled sensitivity-aware caching framework. Specifically, we formalize the caching error through an analysis of the model output sensitivity to perturbations in the denoising inputs, i.e., the noisy latent and the timestep, and show that this sensitivity is a key predictor of caching error. Based on this analysis, we propose Sensitivity-Aware Caching (SenCache), a dynamic caching policy that adaptively selects caching timesteps on a per-sample basis. Our framework provides a theoretical basis for adaptive caching, explains why prior empirical heuristics can be partially effective, and extends them to a dynamic, sample-specific approach. Experiments on Wan 2.1, CogVideoX, and LTX-Video show that SenCache achieves better visual quality than existing caching methods under similar computational budgets.
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。