Skip Navigation

Posts
19
Comments
563
Joined
2 yr. ago

  • Seems more like ollama as approach. More easy but less optimized than llamacpp?

  • Yes, llama swap also give you a pretty web based statistics of all the calls and runs for every model with t/s and more statistics.

    It's pretty neat... You can also load and unload models manually, define groups for models that fit together in vram and so on.

    It gives you that automation that people coming from ollama are used to.

    under the hood all it does is running llama serve. You convert your models.ini to a yaml file 1:1 (plus a few more flexibility).

  • Yes but llama server does not unload models to fit a different model. At least, didn't managed to have just llama.cpp swap between models that fill up my vram... The second one would fail load.

  • Llama.cpp of course! After I did the switch, never looked back. Add llama-swap in front of you switch between models often.

  • You missed my point?

    why can't you just use a wireguard connection between your device and the server, or network?

  • I don't get the need for those tunnels.

    Wouldn't a simple wireguard setup be enough un less you are behind CG-NAT?

    Wireguard it's there, free, open source, doesn't rely on any external server or company...

    What does these "tunnels" actually add?

  • This won't work for Android windows and such devices. And it's a pain to setup, won't work 100% and some scanners are ven worse .

    Scanservejs is the final solution for all platforms.

  • Op want to upload a file to a web GUI and print it from there, not share printers with other computer/ devices.

    Also, scanservejs is the the sane web GUI to actually do the same with scanner.

    For scanning it make sense as there are no shared scanning solutions across platforms. For printing, it's of little use as other pointed out you can print to cups from anything basically.

    But o get the need for a web printing page, as uploading a file and hit print button is a zero setup, and while simple even remote cups printing is not 0% extra setup, you have at the very least to select a printer or hit "share as" and select a print service (android). YMMV

  • 48 disks seems a nightmare in power consumption, failure rate, and overall cabling management....

    What was the total capacity?

  • I use e self made wood cabinet in the attic where also the PV inverter is located. I used Velcro to strip on the wall (brick wall) stuff like switch firewall etc, only to find those collapsed on the server itself every time. No matter which glue or adhesive i used.

    At the end, a strong unbranded Chinese double sided tape was better and still strong after all big western brand failed over time.

    Maybe a mix of heat and uneven surface was the culprit, who knows.

  • Uptime for home users is useless. More important is recovery, reboot and resume (the 3'Rs)

    You server should reboot automatically, storage should restore itself from unsafe shutdown and your services should resume operations like nothing happened.

    This means test your reboot, ensure stuff doesn't break, and so on.

    Failing storage after an hard reboot is the hardest.

    For network, i have two ISP connections and a an auto switch script Incase one goes down. But honestly it's overkill and not needed for 99% of home hosters.

    Buying some kind of UPS like a battery powered power strip is also nice but keep in mind that require maintenance to replace batteries once every few years (long enough to forget) and so be useless. The ups shall only last enough to perform a controlled server shutdown using NUTS.

  • As a suggestion, ditch ollama and setup llama.cpp. it will work fine with openwebui and it's much more efficient. (Unrelated to the ram/swap issue)

  • Does this app run in a browser or on a server?

  • How is this self hosted? Seems a desktop application

  • Cool man... I did the same last year, best decision ever.

    Pretty easy too

  • Or drive a car

  • That's true...

  • I still like it and i have seen projects with much uglier logos made by humans. Nothing against humans or AI either.

  • Xale, the logo is pretty, on topic and quite cute. Is it made with AI? No idea. Does it really matter?

    Give open source projects a break will you?

    If you don't like the logo just say so.

  • LocalLLaMA @sh.itjust.works

    Suggest a model...

  • LocalLLaMA @sh.itjust.works

    Followup on "Some general questions on AI setup"

  • LocalLLaMA @sh.itjust.works

    Some general questions on AI setup

  • LocalLLaMA @sh.itjust.works

    Ok, time to move from Ollama + OpenWebUI

  • Selfhosted @lemmy.world

    LazyNVR: a different approach in hosting IP webcams

    codeberg.org /LazyNVR/lazynvr-sources
  • Selfhosted @lemmy.world

    My personal Simple Dashboard

  • Selfhosted @lemmy.world

    ExcaliDash: self host ExcaliDraw with multi user and server side storage

    github.com /ZimengXiong/ExcaliDash
  • Selfhosted @lemmy.world

    IPv6

  • Selfhosted @lemmy.world

    Spam blocking in 2026

  • LocalLLaMA @sh.itjust.works

    How to... (Maybe I am missing something)

  • Selfhosted @lemmy.world

    Selfhost an LLM

  • Selfhosted @lemmy.world

    Spotify sync web gui

  • Selfhosted @lemmy.world

    Conduwuit is dead, long live Tuwunnel!

    github.com /matrix-construct/tuwunel
  • Selfhosted @lemmy.world

    I found a nice gem

  • Selfhosted @lemmy.world

    Self-hosting minecraft

  • Selfhosted @lemmy.world

    Immich: opinion revised

    wiki.gardiol.org /doku.php
  • Selfhosted @lemmy.world

    Issue with wireguard and advance routing

  • Selfhosted @lemmy.world

    Stalwart mail server

  • Selfhosted @lemmy.world

    First self-hosted post!