Error running on vultr? Or what am i doing wrong??

[bpawnzZ]

I build and all goes fine. but when i try to docker-compose up this i get the following error. Torch is installed and nvidia-smi shows that my drivers are good? Web-Ui runs for about 5 seconds and shuts down and enters a loop doing this until i stop it. I am able to get to port 1111 but cannot connect to web-ui. Any suggestions? Thanks!!

RuntimeError: Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check

[robballantyne]

You'll need the Nvidia container toolkit installed on the VM for the GPU to be available to docker.

The compose file will also need the Nvidia devices section uncommentiing

https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html should have everything you need to get it up and running

[bpawnzZ]

That is not the issue. I have nvidia toolkit installed inside of Vultr. And i tried testing it on a Desktop that has stable diffusion working inside of docker already and I am having the same issues. Basically webui starts and then crashes. When i check services on ip:1111 Jupyter and webui fail. I am really at a loss currently. Tried to do a clean rebuild(none of my personal extensions and models) and same errors. so nothing to do with anything I initially added. Also have tried to compile on more than one machine. Same errors. Is anyone else having similar issues? I am all ears.