template
I'm trying to start a pod in Secure Cloud. When I select an appropriate server, I click Runpod Pytorch 2.8.0 from Change Template and it immediately resets it to 2.1 (which has CUDA 11.8, which won't work for my application.) Why is it doing this and how do I fix it?
8 Replies
is the template selection isn't working?
can you try a different browser
That's correct.
I tried in in Edge. Same problem.
And I forgot to mention, I tried pasting the link to the desired template into the Container Image Field under Edit Template, and that didn't work either
@Bryan Tapscott - RVA GKY
Escalated To Zendesk
The thread has been escalated to Zendesk!
When I start the container with this template it give the following error
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 prestart hook #0: 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
the gpu driver or the gpu itself doesn't support cuda 12.8
https://docs.nvidia.com/deeplearning/cudnn/backend/latest/reference/support-matrix.html#gpu-cuda-toolkit-and-cuda-driver-requirements
probably an old driver problem
This template requires underlying machine running cuda 12.8, make use the cuda filter at the top deploy page.
when selecting a certain cuda version all machines show as unavailable
has anyone faced a similar issue?
nvm lol
try lower cuda versions