Does not work on Runpod

[dakotaplainsllc]

I tried this template on 2-8-2026. This template does not work. I followed the instructions in the readme to change the 5000 to 7860. But none of the UI's work.

[WouterGlorieux]

Did you get an error? I use it everyday and it does work. It just takes a little bit of time until the server is started because it copies the data to the workspace on first startup. In the logs you should see when the server is ready.

[dakotaplainsllc]

I get various various errors about Conda and PIP. Tried to attached the log.

logs.txt

[WouterGlorieux]

Hmm, that is odd, never seen those kinds of errors. What GPU are you using? Are you using an existing network volume?

[dakotaplainsllc]

Yes, using an existing network volume (all it contains is a Llama model I want to use). Using (1) RTX PRO 4500.
I am assuming your template doesn't contain any models by itself.

[WouterGlorieux]

I think you better start with a fresh network volume, then download the model from the UI.

[dakotaplainsllc]

Thanks for the response. It works with models models downloaded through the UI, but seems to error if someone tries to move their own model into the models folder. I am using the Llama 70B Q4 GUFF model. It gives countless errors about LLama.cpp not being present.
Thanks for setting this template up. Seems to me that things are changing so fast that compatibility is hard to maintain.