Hi,
I followed the steps outlined in your blogpost, and my setup works perfectly and is identical to the one described.
For certain applications, I need access to shared libraries such as libnvdla_compiler.so
and libnvdla_runtime.so
during build and runtime. Typically, these are included under /usr/lib/aarch64-linux-gnu/tegra
in the container from the host when --runtime nvidia
is used.
Is there a way to access these libraries from within a container when deploying containers to edge devices using Balena?