LESA: Learnable Stage-Aware Predictors for Diffusion Model Acceleration

arXiv:2602.20497v3 Announce Type: replace-cross Abstract: Diffusion models have achieved remarkable success in image and video generation tasks. However, the high computational demands of Diffusion Transformers (DiTs) pose a significant challenge to their practical deployment. While feature caching is a promising acceleration strategy, existing methods based on simple reusing or training-free forecasting struggle to adapt to the complex, stage-dependent dynamics of the diffusion process, often r