Easy and flexible way to use online LLMs #41
magillos
started this conversation in
Show and tell
Replies: 2 comments
-
|
I uploaded to AUR: https://aur.archlinux.org/packages/llamification-git And I had to: |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
Great! Added a hint to the README https://github.com/manticore-projects/aurscan#backend-fallback-chain |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I wanted be able to connect Aurscan to online LLMs easily and switch them (to run a few test with different models). LiteLLM allows that but it's an overkill, and requires editing config and restarting. It also has history of being compromised, which is makes it a bit ironic to use it for security scans.
So here comes LLamification.
It will let you to use Aurscan with an online provider (Openrouter, Nvidia NIM and others):
-install python-pyqt6 and python-aiohttp (available in official repos)
-set your e. variable with
AURSCAN_OPENAI_URL="http://127.0.0.1:4001/v1/chat/completions"(default port, can be changed),-clone the repo
-run app with
python -m llamification -g, set up your provider and and pick the model, start proxy. That's it.You will be able to change providers and models on the fly.
python -m llamificationwill run it in headless mode with the last used settings.PKGBUILD and AUR package may arrive later.
(The app usage goes beyond Aurscan, of course. It should work whenever you need to disguise your online LLM as Ollama).
Beta Was this translation helpful? Give feedback.
All reactions