Results 3 comments of Cedvano

> @juntaosun You can try our video depth model [ViTA](https://github.com/KexianHust/ViTA) is based on DPT (MiDaS 3.0). Not working, error wint torch.load(checkpoint) and cuda.

Oh, sorry, I don't see this parameter :( Now, that's work fine. Thanks.

It's grey, so from 0 to 255 for displacement. Afterwards, I divide into the number of layers I need