SeiT: Storage-Efficient Vision Training with Tokens Using 1% of Pixel Storage.

Song Park, Sanghyuk Chun, Byeongho Heo, Wonjae Kim, Sangdoo Yun

Browse the full ICCV paper archive.