用最优传输思想优化扩散模型采样缓存，实现高效高质生成。
OTCache: Optimal Transport for Geometry-Aware Caching in Diffusion Models
- 基于最优传输建模缓存策略演化，避免传统方法的假设缺陷。
- 在FLUX.1、Qwen-Image等模型上提速4.5倍以上，质量更优。
- 无需训练，适合追求速度与精度平衡的生成应用开发者。
我们提出OTCache，一种无需训练的扩散模型采样加速框架，通过缓存调度预测实现高效推理。现有基于图的缓存方法依赖加性独立性假设，在低NFE（数值求解步数）场景下常失效。OTCache受最优传输启发，将不同推理预算下的缓存调度建模为策略空间中的平滑演化过程。该框架分三阶段：(1) 在保守预算下使用图基方法获取高保真参考调度；(2) 通过Optuna优化在极端低预算下进行轻量级锚点搜索，目标为端到端感知质量；(3) 利用连续形变表示，通过分位数插值预测目标预算下的调度。在FLUX.1 [dev]、Qwen-Image和HunyuanVideo上的实验表明，OTCache分别实现4.5倍、4.7倍和3.66倍加速，且生成保真度持续优于当前最优缓存基线。本工作为扩散模型加速提供了基于最优传输的新视角。代码：https://github.com/UnicomAI/OTCache
原文摘要 · Abstract (English)
We propose OTCache, a training-free framework for accelerating diffusion sampling via caching schedule prediction. Existing graph-based caching methods reduce redundant computation by optimizing shortest-path objectives, but rely on an additive independence assumption, which often breaks down in the low NFE regime. To address this issue, OTCache models caching schedules across inference budgets as a smooth evolution in policy space, inspired by Optimal Transport (OT). The framework consists of three stages: (1) obtaining a high-fidelity \textbf{reference schedule} using a graph-based caching method under a conservative budget; (2) performing a lightweight anchor search under an extreme low-budget setting via Optuna optimization with an end-to-end perceptual objective; and (3) predicting schedules for target budgets via quantile interpolation between the reference and anchor policies using continuous warping representations. Experiments on FLUX.1 [dev], Qwen-Image, and HunyuanVideo show that OTCache achieves 4.5x, 4.7x, and 3.66x acceleration, respectively, while consistently improving generation fidelity over state-of-the-art caching baselines. This work provides a new perspective on accelerating diffusion models through Optimal-Transport-inspired schedule modeling. Code:https://github.com/UnicomAI/OTCache
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。