SoftCache：扩散Transformer的自适应步长缓存
SoftCache: Adaptive Step Size Caching for Diffusion Transformers
AI总结：
SoftCache是一种无需训练的自适应步长软缓存方法，通过利用陈旧性信号的余量调整缓存步长，在FLUX、HunyuanVideo等模型上实现3.5-6.9倍加速，且在四项保真度指标上优于所有基线。
摘要：
扩散模型和流模型通过迭代步骤生成样本，每一步都需要对大型网络进行完整前向传播，该过程是生成成本的主要来源。无训练缓存通过复用近期评估的激活值而非重新计算来加速这些模型。动态缓存方法通过对网络输入进行廉价探测并累积为陈旧性信号，当信号超过阈值时执行一次新评估以决定何时刷新，这类方法采用硬缓存：信号是刷新与复用之间的二元开关，无论每一步是新计算还是缓存复用，都会按规定步长推进。我们发现该信号包含的信息远不止这个开关：其低于阈值的余量（headroom）可预测复用引发的误差。我们提出SoftCache，采用软缓存：替代二元开关，余量决定每一步缓存步长的推进距离。当余量较大时，采样器会跳过多个预定节点；临近刷新时，步长会收缩回原始步长。这些自适应跳跃可能落在预定节点之间，使采样器脱离固定网格。SoftCache无需训练，也无需添加骨干网络相关设置：在图像生成模型FLUX及视频生成模型HunyuanVideo、Wan2.1上，其加速比达3.5倍至6.9倍，在全部四个保真度指标上均优于所有基线方法。
英文摘要：
Diffusion and flow models generate a sample over iterative steps, each a full pass through a large network, and that pass dominates generation cost. Training-free caching accelerates them by reusing activations from recent evaluations instead of recomputing them. Dynamic caching methods decide when to refresh by accumulating a cheap probe of the network input into a staleness signal and spending a fresh evaluation once it crosses a threshold. They perform hard caching: the signal is a binary switch between refresh and reuse, and every step, fresh or cached, advances by the prescribed step size. We observe that the signal carries much more than this switch. How far it sits below the threshold, its headroom, predicts the error induced by reuse. We propose SoftCache, which performs soft caching: instead of a binary switch, the headroom sets how far each cached step goes. Where the headroom is large, the sampler jumps past several scheduled nodes, and as a refresh nears, its steps shrink back toward the original step. These adaptive jumps may land between the scheduled nodes, taking the sampler off its fixed grid. SoftCache requires no training and adds no per-backbone settings: on FLUX for images and on HunyuanVideo and Wan2.1 for video, at speedups from $3.5\times$ to $6.9\times$, it is ahead of every baseline on all four fidelity metrics.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
