Question about using ground truth for validation

[MingkangXiong]

Thank you for your nice sharing, your code is very nice!

I found that you add "--with-gt" in train_resnet_256.sh. Is that mean using ground truth to validate the depth during training? Is that suitable for unsupervised learning?

[JiawangBian]

Thanks for your attention. The code is based on sfmleaner pytorch, where with-gt is used. However, we find this is important for sfmlearner but not so critical to our method, because sfmlearner is easy to overfit. You may find this in issue #2 , where our method works well without gt for validation, but sfmlearner(named basic) fails.