Closed sconlyshootery closed 3 years ago
For our results on the KITTI Depth Benchmark (Sec. 5.3), the depth prediction is trained alone without other semantic branches -- the dataset does not provide any supervision for those branches, so they are omitted during training/inference.
Thank you, help me a lot.
Thank you for your wonderful work, I just want to know what performance will be if depth prediction is trained alone without other semantic branches?