Is it possible to train the model with KITTI odometry dataset that have sparse GT depth image?
Is it possible to train the model with KITTI odometry dataset that have sparse GT depth image?