TT-VidT: Decoupling the Temporal Axis for Efficient Motion-Centric Video Pretraining
TT-VidT pairs a DINOv3 spatial ViT with a temporal layer, leading motion benchmarks at lower FLOPs.
1 stories, newest first · one card per story
TT-VidT pairs a DINOv3 spatial ViT with a temporal layer, leading motion benchmarks at lower FLOPs.
Today 1.2K readers · 1.4K views · 193 reads · all time 203K views