I’m having Jetson Nano 2GB Dev kit with Jetpack version: r32.5.1. In order to run the container, first I could successfully pull the l4t-tensorflow container. But then when I tried following command: sudo docker run -it --rm --runtime nvidia --network host nvcr.io/nvidia/l4t-tensorflow:r32.5.0-tf2.3-py3py3 to start the interactive session in the container, it just makes me superuser. What should be the expected output or how to start the interactive session in the container so that I can import tensorflow?
After the command, you are already launching the container.
To import TensorFlow, just run the python code on the console directly.
$ python3 >>> import tensorflow >>> tensorflow.__version__
The issue is when we execute the above docker command it goes into root user shell and doesn’t show up any container. How to troubleshoot it?
Basically, we are not getting an interactive session from the container?
I believe what you are referring to is actually the shell into the container. The container terminals appear to be a root shell. If you see a
.dockerenv file in the top-level root directory of the filesystem, then you are inside the container:
# ls -ll /.dockerenv -rwxr-xr-x 1 root root 0 Nov 1 19:34 /.dockerenv
Alternatively, if you can run
python3 -c 'import tensorflow', then you are probably inside the container too (unless you have also installed TensorFlow outside of container on your device, in which case this may be ambiguous)
OK. Thanks I got it. It seems I am inside the container. But, is there a container where TF, Python3 and Jupyter Notebook can be launched? The issue here is I am unable to launch Jupyter notebook from the container.
Hi @rgohain, the l4t-ml container has TF1, Python3, and Jupyter notebook already installed:
Thank you. It works. But, still I have an issue as TF version is 1.15. Is there any container where TF version is 2,0?
You can find some containers for TensorFlow v2.x below:
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.