@Scootter I was somehow able to reproduce your issue on my test server but not on my main server.   This maybe sounds really stupid but to solve the issue I simply copied the GPU UUID from the plugin again in the template at NVIDIA_VISIBLE_DEVICES and clicked Apply again and the container is now running again... This also survives reboots too.   Don't know what caused this, maybe it was the up and downgrades that I've did on my test server...
    • Like
    1