nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.8, please update your drive

error starting container: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running hook #0: error running hook: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy' nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.8, please update your driver to a newer version, or use an earlier cuda container: unknown
2 Replies
Dj
Dj3d ago
You're using a 5090 (or b200) which requires a specific CUDA version you can filter for in the dropdown. It can be a little annoying to have to manually verify the versions, and I'm working on a solution.
Jason
Jason3d ago
In this case the cuda in the machine is older than 12.8,so try to use cuda versions(container image) like 12.4 for now Except you're using 5090 or b200 You can filter out the cuda version in your endpoint

Did you find this page helpful?