[ICCV2023 Oral] Unmasked Teacher: Towards Training-Efficient Video Foundation Models - View it on GitHub