Author name: Yongsheng Yu, Wei Xiong, Weili Nie, Yichen Sheng, Shiqiu Liu, Jiebo Luo

arXiv:2511.20645v2 Announce Type: replace
Abstract: Latent-space modeling has been the standard for Diffusion Transformers (DiTs). However, it relies on a two-stage pipeline where the pretrained autoencoder introduces lossy reconstruction, leading to …