Jetson Nano docker image for cross-compilation

Hi, I would like to cross-compile packages such as OpenCV 4.8 with GPU and Cuda support to use in Jetson Nano. I started creating docker containers to do this from l4t-base and similar, but these containers as far as I know don’t have cuda, cudnn installed on the image, they have links to the libraries that are located in the host machine. This makes it much more difficult to cross compile using a docker image.

Please, is there any docker image available that would make it possible to cross-compile packages that depend on cuda, cudnn, etc. on a x86 machine? Should I be looking for a different approach than docker containers?

Thanks in advance for your help!

Hi,

We do have a cross-compiling container for JetPack 5.

For Nano, which depends on JetPack 4, please refer to the following doc to setup the environment:
https://docs.nvidia.com/vpi/1.2/sample_cross_aarch64.html

Thanks.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.