dontLoveBugs / DORN_pytorch

PyTorch implementation of Deep Ordinal Regression Network for Monocular Depth Estimation
299 stars 67 forks source link

How to deal with the loss backward function? #25

Closed dongda118 closed 4 years ago

dongda118 commented 4 years ago

Hi, is anyone melt with the problem about the loss.backward() function? in the original paper, it need calculated with x(w,h) and Pk, but i don't know which layer's output is the x(w,h) ? I don't find the code ,can you or anyone tell me how to deal with it? Thanks very much. and I hope to receive your answers.