Closed shuther closed 1 month ago
Hi @shuther,
Can you please provide the full details of what you are trying to achieve, a bit more details about your machine (like what OS you are using) and what you have already tried?
OS: VM linux I am trying to leverage ollama/litellm openAI compatible endpoints to test further data-to-paper. It means I need to enforce the openai base url. I tried changing env.py (to update the openai base url as well as the deepinfra baseurl) to set a different endpoint but it is as if only gpt4 was called. I have no logs on my endpoint side (that it tried to connect to).
Thanks @shuther, for giving more details.
I have created a new branch (attached to this issue), that should support llama3 (as an example), hopefully it will also make it more clear how to generalize to other ollama/litellm models. Specifically, look at the commit c699eebe04870a787f18ed732da2cd227e8e21a1.
even with updating env.py, it seems the script doesn't want to use the set openai base url: script used: