Abstract: Video pose transformers (VPTs) have demonstrated remarkable performance in 3D human pose prediction. However, transformer-based architectures are often computationally intensive, leading to ...