Closed katsunori-waragai closed 3 months ago
All I have to do is to execute as follows.
cd /root/TranspareNet/grnet_point_cloud_completion/extensions/gridding
python3 setup.py install
cd /root/TranspareNet/grnet_point_cloud_completion/extensions/gridding_loss/
python3 setup.py install
cd /root/TranspareNet//grnet_point_cloud_completion/extensions/cubic_feature_sampling/
python3 setup.py install
cd /root/TranspareNet/grnet_point_cloud_completion/extensions/chamfer_dist/
python3 setup.py install
resolved
Thank you for your implementation of TranspareNet. I found it very interesting as a library to improve the depth from RGB and raw Depth images, including transparent objects. I would like to port it to Jetson AGX Orin and make it work. So I built the environment. However, as the module for gridding has not been created I could not run inference_transparenet.py. How can I create the gridding module?
Environment
Jetson AGX Orin
Ubuntu 20.04 LTS JetPack5.1, L4T 35.2.1 CuDNN 8.6.0, CUDA 11.4.19, TensorRT 8.5.2,
Docker environment
python3 -m pip install -r requierments.txt
Error
My understanding of the problem situation
I assume that the cu files and cpp files in the following locations are not compiled and not bound to python3, which is causing the problem.
Expected response
I would like to see a way of binding to python3 using cu files, cpp files.