[Section request] AI/LLM

[C-Fu]

Application Name

Ollama, Stable Diffusion, LocalAI, gpt4all, etc

Website

Description

I believe a lot of the readily-available LXC-based scripts are already using (selfhosted) AI tools, so might as well we start having an AI section for it.

Due Diligence

• I have searched existing scripts and found no duplicates.
• I have searched existing discussions and found no duplicate requests.
• This is not a game-related request.

[MickLesk]

Thats are more then 1 Request.
Ollama und OpenWebUI are available

[michelroegl-brunner]

I think he is refering to the frontend/website the, not the install Scripts. I guess he is taking about a "AI" Category.
I think this is not a bad idea. @BramSuurdje (Maybe you can have a look at it)

[MickLesk]

When we have enough Scripts yes

[darknessblade]

For things like Stable-Diffusion, I do think there need to be multiple helper-scripts [or force users to go trough the manual setup script]

since not everyone runs it on supported hardware. since for stable-diffusion a NVIDIA GPU is recommended.
There is a AMD fork, but its quite broken and doesn't work well with certain features like upscaling
[at least under windows], due unclear or incomplete documentation that is outdated.

For Stable-diffusion, the helper-script should not have a "AUTO-install" option,
It should force people to go trough the manual prompt where they need to specify their hardware specifications for SD/ the AI tools they want to install.
It should also give ample warnings if unsupported hardware is detected.

A example for me would be: I have proxmox running on a Intel N100. and would be able to run a StableDiffusion fork called "easydiffusion" in a LXC container,
The IT/S won't be as high as with GPU rendering. but it should still be more than enough for simple Image generation.

[spijkey87]

I would like this