nvidia jetson docker

I am using the nVidia Jetson AGX Xavier edge device and have pulled a Docker image from a source. I started a container and updated some Python third-party libraries within the container. However, when I reopen a container using the following command (sudo docker run -t -i --runtime nvidia fedml/fedml:latest-nvidia-jetson-l4t-ml-r35.1.0-py3 /bin/bash), the updated third-party Python libraries revert back to their original versions. I am quite frustrated. Is there any way to keep the updated state of the libraries?

Yes, containers are discarded all the time.

Instead you should create your own image with a Dockerfile.