Runtime failure when attempting to run pytorch docker

Hello,
I am relatively new to running docker containers. I have run a few before with explicit directions online however didn’t really know all of what I was doing. I am attempting to build a yolo v5 model on a jetson xavier nx which to do that I have seen where people use a pytorch docker container first to build ontop of. Not really sure all of what that does for me but would appreciate any info one might offer.

Anyway, after pulling the latest pytorch container (nvcr.io/nvidia/l4t-pytorch r35.1.0-pth1.13-py3 3ad4d95c8abe 8 months ago 11.7GB) for my specific jetpack build (Package: nvidia-jetpack ; Version: 5.0.2-b231), I see that the previous pytorch container that I pulled upon creation of this hard drive has been updated. Unfortunately, when I go to run this container I am met with the following error message:

sudo docker run --gpus all -it --rm -v /home/famulus/data:/data/ nvcr.io/nvidia/l4t-pytorch:r35.1.0-pth1.13-py3
docker: Error response from daemon: failed to create shim task: OCI runtime create failed: failed to create NVIDIA Container Runtime: failed to construct OCI spec modifier: failed to construct discoverer: failed to create Xorg discoverer: failed to locate libcuda.so: pattern libcuda.so...* not found: unknown.

Any guidance would be much appreciated, I would like to get this running soon so that I can help a friend out to ensure a smooth running experiment of theirs.
Thank you!

1 Like

errors may occur due to missing dependencies or incorrect installation of packages. It is recommended to carefully review the installation instructions and ensure that all necessary dependencies are installed correctly.