The size of this community alone shows the need for it. That said, feddit.org seems like a good new home for it since .org is a language neutral tld and the server is german AND English speaking after all.

That Kalium with two bonds tho

Glad i could help ;)

You can get different results, sometimes better sometimes worse, most of the time differently phrased (e.g. the gemma models by google like to do bulletlists and sometimes tell me where they got that information from). There are models specifically trained / finetuned for different tasks (mostly coding, but also writing stories, answering medical questions, telling me what is on a picture, speaking different languages, running on smaller / bigger hardware, etc.). Have a look at ollamas library of models which is outright tiny compared to e.g. huggingface.

Also, i don’t trust OpenAI and others to be confidential with company data or explicit code snippets from work i feed them.

If you’re lucky you just set it to the wrong version, mine uses 10.3.0 (see below).

I tried running the docker container first as well but gave up since there are seperate versions for cuda and rocm which comes packaged with this as well and therefor gets unnecessary big.

I am running it on Fedora natively. I installed it with the setup script from the top of the docs:

curl -fsSL https://ollama.com/install.sh | sh

After that i created a service file (also stated in the linked docs) so that it starts at boot time (so i can just boot my pc and forget it without needing to login).

The crucial part for the GPU in question (RX 6700XT) was this line under the [service] section:

Environment="HSA_OVERRIDE_GFX_VERSION=10.3.0"

As you stated, this sets the environment variable for rocm. Also to be able to reach it from outside of localhost (for my server):

Environment="OLLAMA_HOST=0.0.0.0"

I have my gaming pc running as ollama host when i need it (RX 6700XT with rocm doing the heavy lifting). PC idles at ~50W and draws up to 200W when generating an answer. It is plenty fast though.

My mini pc home server is running openwebui with access to this “ollama instance” but also OpenAIs api when i just need a quick answer and therefor don’t turn on my pc.

It’s a reference to the latest fireship video… I think:

https://youtu.be/AdygBbbEnco

Anna’s archive would be the go to i think. You can choose the language in the sidebar.

Is that Florida in your atlantic ocean or are you just happy to see me?

I hope it went well :) i was completely ready to go back changing the image tag to v2 but didn’t need to.

Alpine

I think the whole “men don’t know where the clitoris is” in reality means either

• they don’t know what to do with it
• they don’t care

Finally. Hail the google overlords!

Bei meiner Fritzbox kann man unter Freigabe ein Gerät aussuchen (raspi), und weiter unten dann die ports mappen. Heißt du suchst dir einen externen port aus, über den du dich dann mit der Fritzbox verbindest (was auch immer der speedport ihr für eine IP gegeben hat + der Port den du ausgesucht hast) und einen internen port / “port am gerät” (der am raspi den du freigeben möchtest, bspw. 8096 für jellyfin)

VPN am speedport einrichten falls das geht (bei der fritzbox geht das ja) und Port forwarding an der Fritzbox auf den Raspi.

Oder VPN auf dem raspi und zwei mal port weiterreichen (am speedport und an der Fritzbox), halte ich aber für umständlich. Quelle: hab sowas ähnliches mal gemacht (fritzbox - openwrt - raspi).

Oder “Modernerer” Ansatz: tailscale, zerotier, twingate (wähle eins oder probier alle mal durch :D). Kannst du auf deinem raspi installieren und auf dem client deiner wahl. Benötigt aber einen externen Dienst um die Geräte zueinander zu vermitteln.

Bricked my pc twice because of the bootloader and couldn’t repair it. From now on i just nuke my system if something is fucky and have a shell script do the installing of packages etc.

I had a friend come over to my place to fix her laptops wifi. After about an hour searching for any setting in windows that i could have missed, i coincidentally found a forum where one pointed out this could be due to a hardware wifi switch…

Bis Tiktok mal was tut muss viel passieren. Ich würde sogar behaupten (und das als linksversiffter der die AfD abgrundtief hasst) dass es noch deutlich schlimmer mit anzusehende rabbit holes dort gibt.

Trotzdem ein Schritt in die richtige Richtung.

I’m not sure if I understand your question, but if you are trying to build a solution, you will have to know the problem. I am writing said mailings, database, rest interfaces etc. for which Java and Spring boot is pretty useful. Some people might consider this antiquated. I also used python for data science stuff in the past. Neither would i like to have been using java back then, or python now. But in both cases i needed to know what i was building before i (or my employer) chose the technology to use.

If you are offered a position where you will have to use a technology for something you think its not a good fit for then run. Some people might even be more interested in you if you tell them as it makes you look more experienced. And you’re right, it is usually not worth it to have religious wars over 0.xy percents of performance gain, as long as you’re not trying to build a house with a screwdriver.

Im so looking forward to this. When i tried to use tmpfs / ramdisk, the transcoding would simply stop because there was no space left.