1 paper
Junwen Tan, Jinglin Liang, Hongyuan Chen +1
Though rectified flow models have achieved remarkable performance in image, video, and 3D generation, their practical deployments are challenged by slow inference speeds. Prior acc…