Significant increase of docker image size for R34.X comparing to R32.X

I flashed my Jetson AGX Xavier with T4L R.34, and noticed, that related docker image for inference is 5x bigger than for R.32.

What is the reason of that?
I do not have any extra memory to add currently and that prevents me from using prebuilt, dockerized solution for inference, as I have 6GB of memory left.
Or am I missing something and there is workaround?

Would be grateful for any advice :)

In R32, the libs are mounted from the host to the containers, but in R34, the container includes the actual libraries. This gives the benefit of running different lib versions in the container, eg: host runs R34 but container runs R35.


Just as user100090 mentioned.
From JetPack 5.0DP, we include the libraries instead of mounting from the host.
So the size of the container will increase.

Since we don’t mount libraries anymore, you don’t need to have the libraries installed on the Jetson host.

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