EVEREST: Efficient Masked Video Autoencoder by Removing Redundant Spatiotemporal Tokens.

Sunil Hwang, Jaehong Yoon, Youngwan Lee, Sung Ju Hwang

Browse the full ICML paper archive.