根据特征动态自适应选择加速策略，实现扩散模型无损提速。
Let Features Decide Their Own Solvers: Hybrid Feature Caching for Diffusion Transformers
- 按特征维度差异设计混合微分方程求解器，实现差异化缓存。
- 在多个模型上实现5.55倍以上加速，无需重新训练。
- 适合追求高效生成的视觉大模型应用者。
扩散变压器在图像和视频生成中达到顶尖保真度，但其迭代采样过程因每步高成本的Transformer前向计算成为主要瓶颈。为缓解此问题，特征缓存作为无需训练的加速技术应运而生，通过重用或预测隐藏表示。然而，现有方法通常对所有特征维度采用统一缓存策略，忽略了其异构动态行为。为此，本文将隐藏特征演化建模为跨维度的混合常微分方程（ODE），提出HyCa——一种受混合ODE启发的自适应缓存框架，实现维度级缓存策略。HyCa在多种领域与模型上均实现近乎无损加速：FLUX上达5.55倍、HunyuanVideo上5.56倍、Qwen-Image与Qwen-Image-Edit上均达6.24倍，且无需重新训练。
原文摘要 · Abstract (English)
Diffusion Transformers offer state-of-the-art fidelity in image and video synthesis, but their iterative sampling process remains a major bottleneck due to the high cost of transformer forward passes at each timestep. To mitigate this, feature caching has emerged as a training-free acceleration technique that reuses or forecasts hidden representations. However, existing methods often apply a uniform caching strategy across all feature dimensions, ignoring their heterogeneous dynamic behaviors. Therefore, we adopt a new perspective by modeling hidden feature evolution as a mixture of ODEs across dimensions, and introduce HyCa, a Hybrid ODE solver inspired caching framework that applies dimension-wise caching strategies. HyCa achieves near-lossless acceleration across diverse domains and models, including 5.55 times speedup on FLUX, 5.56 times speedup on HunyuanVideo, 6.24 times speedup on Qwen-Image and Qwen-Image-Edit without retraining.
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。