SeiT++: Masked Token Modeling Improves Storage-Efficient Training.

Minhyun Lee, Song Park, Byeongho Heo, Dongyoon Han, Hyunjung Shim

Browse the full ECCV paper archive.