Fetching the paper…

UniFormerV2: Spatiotemporal Learning by Arming Image ViTs with Video UniFormer · Around