[CVPR'22 Oral] TTVSR: Learning Trajectory-Aware Transformer for Video Super-Resolution - View it on GitHub