why the disp_gradient_loss and lr_loss is zero, when training on 3D movies?

mrharicot / monodepth

Unsupervised single image depth prediction with CNNs

Other

2.21k stars 628 forks source link

Open redcping opened 5 years ago

stackoverflow2 commented 5 years ago

We have train this module on 10 3d movies still we are getting the same issue also the total loss is not getting converging.

zouzhekang commented 4 years ago

I transform it into tensorflow 2.0 version and I meet this issue too ...have you resolved it?