aiqbench analysis
EpaCache addresses the high inference costs of diffusion models by introducing a training-free policy that intelligently manages cache reuse to minimize error propagation across timesteps. This approach significantly improves the latency-fidelity trade-off for both image and video synthesis, achieving substantial speedups without compromising output quality on leading open-source models.
EpaCache: Error-Propagation-Aware Caching for Accelerating Diffusion-Based Visual Generation
arXiv:2608.29264v1 Announce Type: new Abstract: Diffusion-based visual generative models deliver strong image and video synthesis quality but incur high inference costs because sequential samplers repeatedly evaluate large networks. Caching-based methods reduce inference latency by reusing intermediate computations across adjacent timesteps. However, existing cache controllers rely primarily on local temporal variation and overlook the trajectory-level consequences of cache reuse. We introduce Error-Propagation-Aware Cache (EpaCache), a training-free caching policy that adaptively allocates the reuse budget on timesteps with lower downstream impact. Experiments on image and video synthesis models demonstrate that EpaCache consistently improves the latency--fidelity trade-off over existing caching methods. On FLUX.1-dev, EpaCache outperforms the prior state-of-the-art caching method in both latency and fidelity, reducing inference time from $11.7$ s to $11.3$ s while improving PSNR from $21.4$ to $22.8$. On HunyuanVideo, EpaCache achieves a $2.63\times$ speedup over uncached inference and improves SSIM from $0.891$ to $0.905$ over the prior state-of-the-art method at matched latency.
Abstract:Diffusion-based visual generative models deliver strong image and video synthesis quality but incur high inference costs because sequential samplers repeatedly evaluate large networks. Caching-based methods reduce inference latency by reusing intermediate computations across adjacent timesteps. However, existing cache controllers rely primarily on local temporal variation and overlook the trajectory-level consequences of cache reuse. We introduce Error-Propagation-Aware Cache (EpaCache), a training-free caching policy that adaptively allocates the reuse budget on timesteps with lower downstream impact. Experiments on image and video synthesis models demonstrate that EpaCache consistently improves the latency--fidelity trade-off over existing caching methods. On FLUX.1-dev, EpaCache outperforms the prior state-of-the-art caching method in both latency and fidelity, reducing inference time from $11.7$ s to $11.3$ s while improving PSNR from $21.4$ to $22.8$. On HunyuanVideo, EpaCache achieves a $2.63\times$ speedup over uncached inference and improves SSIM from $0.891$ to $0.905$ over the prior state-of-the-art method at matched latency.
Artificial Intelligence (cs.AI)
arXiv:2608.29264 [cs.AI]
(or arXiv:2608.29264v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.29264 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 242 words
Related coverage
Building Blocks for Foundation Model Training and Inference on AWS
Hugging Face · May 11, 2026
SenCache: Accelerating Diffusion Model Inference via Sensitivity-Aware Caching
arXiv · Sep 22, 2026
Triton-Based Optimization of Video Sparse Attention on ROCm
AMD · Jul 13, 2026
Informe de riesgos de la IA de frontera (febrero–marzo de 2026)
METR · May 19, 2026