arXivJiarui Chen, Zeqiang Lai, Jiangshan Wang, Ziheng Ouyang, Ye Huang, Xiangyu Yue, Cewu Lu, Chunchao GuoMon, Oct 5, 2026, 10:51 AM PDT
score 17.0
New Method Speeds Up AI Video Generation Without Losing Quality
Original: MC-Sparse: Deconstructing and Closing the Dense-Sparse Attention Gap in Diffusion Transformers
Source: arxiv.org ↗
Writing ELI5 summary…