SparseFormer: Sparse Visual Recognition via Limited Latent Tokens.

Ziteng Gao, Zhan Tong, Limin Wang, Mike Zheng Shou

Browse the full ICLR paper archive.