Add New LXC: Ollama - Githubissues

tteck / Proxmox

Proxmox VE Helper-Scripts

https://Helper-Scripts.com

MIT License

11.68k stars 1.81k forks source link

Add New LXC: Ollama #3201

Closed ulmentflam closed 2 weeks ago

ulmentflam commented 2 weeks ago

Description

This PR adds Ollama as an LXC. Ollama is a self-hosted API for running large language models. It supports the open Llama Models, Phi, Minstral, Gemma, and more. This LXC benefits from GPU passthrough and installation of GPU drivers.

Type of change

[x] New Script (Develop a new script or set of scripts that are fully functional and thoroughly tested)
[x] I have performed a self-review of my code, adhering to established codebase patterns and conventions.
[x] This change requires a documentation update

tteck commented 2 weeks ago

You have not adhered to established codebase patterns and conventions.
This will be a nightmare with the way it's installing NVIDIA and AMD drivers "WSL2 only supports GPUs via nvidia passthrough" No thank you.
Executing https://ollama.com/install.sh inside another bash script is frowned upon.

Just not interested, sorry.

ulmentflam commented 2 weeks ago

Okay, No problem. I apologize for not adhering to the conventions.

I could have also gone through the manual install steps from https://github.com/ollama/ollama/blob/main/docs/linux.md#manual-install which just installs the binary and doesn't install the Nvidia or AMD drivers and leaves the WSL2 passthrough up to the installer.

Out of curiosity what conventions did I miss besides using Ollama's install.sh, script inside of another bash script?

tteck commented 2 weeks ago

header_info

ulmentflam commented 2 weeks ago

I see, you use the slant ASCII font https://github.com/tteck/Proxmox/pull/2605#issuecomment-1979668948. Thanks!

Also thanks for maintaining this, it is fantastic!