Caenorst

Hi @mustafamirrAR , assuming you are talking about the metadata output by the OV extension. assets transforms are the transforms applied to the mesh (things like transform the mesh to...

Hi @LvisRoot , what are you trying to resolve? I can see some `backend_name` in the code base but all the constructors use `backend`: - from_camera_pose: https://github.com/NVIDIAGameWorks/kaolin/blob/master/kaolin/render/camera/extrinsics.py#L278 - from_lookat: https://github.com/NVIDIAGameWorks/kaolin/blob/master/kaolin/render/camera/extrinsics.py#L331...

Hi, can you check your python version? We currently only support python from 3.7 to 3.9 (will support 3.10 in the next release). Please open a separate issue for the...

Hi @HaoZhang990127 , the rasterization is not normalizing the values that are interpolated, how have you produced the `face_vertices_camera` values?

the `face_vertices_camera` is only post extrinsics, so the depth you are seeing is not normalized. Do mind that the current depth that you are display is the distance to camera...

Can you try to move the camera away from the model ? you should see the depth increase

Hi @YIEGR , the metadata are generated by the [Omniverse App](https://developer.nvidia.com/kaolin#kaolin-app). I can help you generating metadata file. To answer your questions: You said you want to use a single...

Hi @kenorb , usually this kind of error happens when the torch version (including the cuda version) is not matching with the wheels you are selecting, can you verify that...

Hi @samiede , please open a new issue, I'll try to help you. Again usually this kind of errors happen when there is a difference in torch version and kaolin's...

Hi @wangsen1312 what is your error message, what is your torch version, what it your cuda version, what is your command-line?