More indepth instructions for dumb idiots like me
SkiddyDiddy opened this issue · comments
Mostly just the title. I really wanna try this cause it seems really cool but i can't really figure out how to do the uvicorn part. I've cloned the repo to a folder with git but i uvicorn elludes me. python, cmd, and miniconda3 all give a generic "not recognized" command when typing uvicorn. I knows its installed i can see it in my python folders i just don't know to use it :c
Also would this work with local models? Like running my own llm model through Text Generation Webui and using that api link instead of openai's api link. You can edit system prompts, termperature, cfg all dat stuff through Text Gen Webui
pip install uvicorn
if the webui supports openai spec, sure, if no then no
try to call uvicorn through python, like python -m uvicorn server:app
Using your own llm "works". It'll generate stuff most of the time but its kind of screwy. I am using LM Studio with gemma.