13 Replies
What does "no execute" mean?
What do the logs say?
The 1st one executed nothing I canceled it manually, the 2nd one canceled it after a timeout and it looks like it was not assigned to a GPU
Where is the message that made it look like it was not assigned to a gpu?
Copy texts, not screenshots please
whats your base docker image?
is that the full logs? yeah i see no "cuda" or "pytorch" signs meaning the gpu can be not loaded
Here's the full log, docker image templates id 98y63mu2ic, multiple wokers are allowed at the same time and only this worker has problems
Try to contact support via the website with your endpoint id
contact button on the left
Its probably issue with the worker itself, not a RunPod issue
I would log a Github isssue on the repo you're using before contacting RunPod support
Is the process that complicated? I've promoted runpod to countless people before because I thought it had prompt customer service, but why is the process so complicated
It could be many possibility thats causing that issue, best to check workers code too but the way the current way of contacting official staffs is my email or ticket
Its definitely the worker code if 1 single person has an issue
If multiple people have issues, then it can be a RunPod issue
Yep