Guan Jie
Guan Jie
Hi, any update on this?
Hi, @fkavache . I am trying to compile this cpp file only. My camera_calibration.cpp is the same as the one in this repo. https://github.com/microsoft/Azure-Kinect-Sensor-SDK/tree/develop/examples/calibration
@fkavache I am using Ubuntu 16 and I only build from source the whole repo but I didn't install it into system directory. If I use cmake, it fails. Is...
hi @naurril . I try to disable tools.caching.on in the main.py, so far the issue is gone.
Thanks for the suggestion. Would give it a try later. In the meantime, did you get a reasonable depth map on your side? May I have a look at it?
Hi, thanks for the visualization. May I know how you got the depth value? Are u using LSS transform for depth prediction as well?
I see, thanks for the suggestion. I am quite confused because although the depth predicted by lss transform is not logical but it works with 3d object detection... Any clue...
I remember in one of the issue the author mentioned the same observation. I am quite curious how accurate is the depth map trained using AwareBEVDepth
Hi. I am solving the same issue. Did you found anything? From the conversation in other issue, I also found that you are having problem training the camera only model,...
Hi. Thanks for the reply. ``` def forward(self, img, depth): B, N, C, H, W = img.size() img = img.view(B * N, C, H, W) feat = self.model.encoders.camera.backbone(img) # feat...