Jetson Orin Nano (JetPack 5.x): torchvision C++ ops (_C) fail to load with PyTorch 2.0.0+nv23.05

AI & Data Science Deep Learning (Training & Inference) TensorRT
4

Hi @AnnaT ,

You can find the TorchVision installation steps in the below topic:
https://forums.developer.nvidia.com/t/pytorch-for-jetson-version-1-11-now-available/72048

Installation

torchvision

If you want to build from source: PyTorch for Jetson

Build From Source

Basically, You just needs to set that one flag (e.g., export TORCH_CUDA_ARCH_LIST="8.7" ), and the compilation on the device will work perfectly. Please do refer to above mentioned KBs.

Thanks.