Haoshen Yu

OnlineCache: Learning Dynamic Caching Policies with Error Correction for Efficient Diffusion Inference

Zhikang Xie, Xichen Ye, Yifan Wu, Haoshen Yu, Li chenan, Peizhu Gong, et al.

Diffusion models have revolutionized generative tasks but incur high latency due to iterative denoising. While cache-based strategies accelerate inference by reusing intermediate features, they largely rely on static, sample-agnostic schedules. We argue that this rigidity overloo…Read more