My team just bought an NVIDIA TX2 to perform object detection, and I would like to give a try to TensorRT for deployment.
I’ve successfully flashed the OS and installed CUDA and cudNN on the TX2, along with tensor flow Object Detection API, but I’m getting trouble installing TensorRT. According to Jetpack 3.2 log during installation, TensorRT is correctly installed, however, I can’t import it in any python script. I also tried to copy the .deb file for installation an the TX2 and run
sudo dpkg -i nv-tensorrt-repo-ubuntu1604-ga-cuda9.0-trt3.0.4-20180208_1-1_arm64.deb
but it can’t complete the installation.