Closed happysujin closed 6 months ago
should I use groundtruth kp2d, cam_vel, cam_poses, pose ... for train dataset? Or make it by processing ViTPose&SLAM?
should I use groundtruth kp2d, cam_vel, cam_poses, pose ... for train dataset? Or make it by processing ViTPose&SLAM?
At first, let's just make using groundtruth.
one for training, one for groundtruth?