[NeurIPS'22] VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training
Primary LanguageJupyter NotebookApache License 2.0Apache-2.0