Generative Video Transformer: Can Objects be the Words?
Yi-Fu Wu 1 Jaesik Yoon 1 2 Sungjin Ahn 1 3
Abstract interest to develop an analogous generative pre-training pro-
Transformers have been successful for many natu- cedure for videos, the computational overhead in dealing
ral language processing tasks. However, applying with videos has made this a difficult endeavor.
transformers to the video domain for tasks such The main ...


雷达卡




京公网安备 11010802022788号







