Wang, S. et al. 2026. PipeDiT: Accelerating Diffusion Transformers in Video Generation with Task Pipelining and Model Decoupling. Proceedings of the AAAI Conference on Artificial Intelligence. 40, 12 (Mar. 2026), 10092–10100. DOI:https://doi.org/10.1609/aaai.v40i12.37976.