Zach Teed

 This picture might help. During training we operate in the local mode, where only the depth for a single keyframe is estimated, this is sufficent to estimate the pose...

Hi, yes during training we only predict the depth for a keyframe (taken to be the first frame in the sequence). However, with more GPU memory, or a smaller batch...

Hi KP, your understanding is correct, we do use RGB-D images for training. When we say is is trained on monocular data, what we mean is that the network is...

It's normal to occasionally encounter singular matrices during training. The training code detects when these occur and will not compute gradients through the system, so this shouldn't be the cause...

Could you provide more information, e.g. what training command did you run, a copy of the tensorboard log file. Under the normal training settings you should get a clean epe...

It looks like you are plotting the translation component of the camera poses. The poses tell how 3D points get mapped to the image. To get the world coordinates of...

You can get a scale correction by doing ```scale = np.sum(gtruth_xyz * pred_xyz)/np.sum(pred_xyz ** 2)``` if gtruth_xyz and pred_xyz are the (x, y, z) predicted/gt coordinates of the camera over...

Looks like cuda error, I don't think batch size should matter in this case. What GPU are you using?

Hi, I will look into this in the next few days

Hi, sorry for the late response. All comments you referenced were correct at the time when they were written. In the initial release of our paper / code (March 2020),...