this post was submitted on 15 Nov 2023
8 points (100.0% liked)

Self-Hosted Main

515 readers
1 users here now

A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.

For Example

We welcome posts that include suggestions for good self-hosted alternatives to popular online services, how they are better, or how they give back control of your data. Also include hints and tips for less technical readers.

Useful Lists

founded 1 year ago
MODERATORS
communick@selfhosted.forum
8
Best GPUs for self-hosted AI? (alien.top)
submitted 1 year ago by Trustworthy_Fartzzz@alien.top to c/main@selfhosted.forum
10 comments fedilink hide all child comments
 

Hello friends,

I’m pretty deep into self-hosting - especially on the home automation side. I’ve got a couple of options for self-hosted AI, but I don’t think they’ll meet my long term goals:

  • Coral TPUs: I have 2x processing my Frigate data. These seem fine for that purpose, but not useful for generative AIs?

  • Jetson Nano: Near as I can tell nothing supports these things except DeepStack, which appears to be abandoned. Bummed these haven’t gotten broader support in the community.

I’ve got plenty of rack space and my day job is managing thousands of machines, so not afraid of a more technical setup.

The used NVIDIA rack mounted Tesla GPU servers look interesting. What are y’all using?

Requirements:

  • Rack mounted
  • Supports local LLM and GenAI
  • Linux-based
  • Works with Docker
you are viewing a single comment's thread
view the rest of the comments
[–] Trustworthy_Fartzzz@alien.top 1 points 1 year ago

I totally agree on the Coral TPUs. Great for Frigate, but not much else. I’ve got 2x of the USB ones cranking on half a dozen 4K stream - works wonderfully.

And I agree in theory these Nanos should be great for all sorts of stuff, but nothing supports them. Everything I’ve seen is custom one offs outside of DeepStake (though CodeProject.AI purports there’s someone now working on a Nano port).

Sounding like a decent gaming GPU and a 2-3U box is the ticket here.