Do you host your own ML / AI / LLM? What do you use, and what do you use it for?
You must log in or # to comment.
Yes. Openwebui/ollama for LLM, comfyui for stable diffusion. I just dick around with it as a toy.
I was put off by ComfyUI, seems awfully complex. How is your experience?
Any suggestions to start? I have Fooocus installed now
It is difficult to understand in the beginning but has great support for premade workflows. It even saves the workflow into its output images so you can drag and drop them into the webui to duplicate the setup that generated the image. Use the internet to get premade workflows and mess around with them to see what the options do and you’ll slowly learn how it works. If you don’t care about precise control over the generations or understanding how image generators works then just use something else more all-in-one.
Same, toyed with it for creating stupid things like bot for telegram, that basically was a 3rd-person NSFW storyteller in RP chat. Sadly, after I made said bot I remember that I don’t have friends to RP with.
That being said, ollama+openwebui kinda sucks: openwebui have “wider scope” and features that you don’t need like auth via social providers and managing multiple accounts, while ollama itself does the opposite and lacks certain features (like proper mmap support to load big models), slow in comparison to pure llama.cpp and generally easily replaceable with lm studio, that provides both - client and server. So yeah, my advise for anyone who want to try it localy - just use lm manager.
Yeah, I’m using qwen 31b a3b on an amd 9070xt requires a bit of cpu offloading, but still plenty fast. Using it wall llama.cpp. Combine that with some mcp’s such as ddg-search to make it truly useful by actually being able to search online.
I mostly use it for small tedious tasks with well defined inputs and outputs. For example when hyprland recently changed from their own configuration language to lua. At first I started going line by line translating my config to the new lua language until I realized oh wait this is exactly the type of thing that ML is useful for. Going from the well defined hyprland configuration language to their also well defined lua syntax. It banged it out in less than a minute with only a single mistake which I easily fixed. The mistake it made was that it forgot to translate the comments to lua. It did it in less than a minute and worked first try. Where as I had made several typos and gotten a few lines wrong when I was doing it by hand.
Not to say that I couldn’t do it. I would have gotten it done in about half an hour, but less than a minute is a lot faster.
I also used it to transform a bunch of unstructured data into json data, so that I could then use purpose built tools like jq to parse that. If I’m having trouble finding certain information. I’ll ask it to find me some resources to look at.
Basically small well defined tasks and parsing data is what I use it for and it seems to be pretty good at that.
What I don’t like is the way companies try to market it to people. I don’t believe people should be trying to summarize emails or messages from loved ones, writing essays or any other creative tasks for the most part. Translating is okay. I don’t expect a machine to be able to decide things for me or to be some filter between me and others.
Nope.
I do, but I am becoming increasingly more disappointed as time goes on. Not just self hosted, llms in general. They sometimes help, but they mislead so many times and waste time that you don’t even notice. I think that’s the trap. When you succeed at a task, you become impressed but don’t notice how many times it failed doing a simple task. And as soon as you scratch the surface, you see how you would have done it differently and perhaps in a better way. Even just googling is bad. It does research for you, but it has no critical thinking and can’t decide what is better from the results it gets (other than google ranking) so it often leads you to think it did as good as you would, when it’s nowhere near as good. Every time I did the googling myself after it did, I did it much better. And I mean MUCH better. Ask it to find the app, it misses the most important ones, hallucinates a bunch, for ex. I found this to be the case with frontier models as well.
Self hosting has its benefits, but seeing how the ecosystem looks right now, concluding this is a huge bubble is inevitable. It reminds me of crypto so much. It looks rich and plentiful, but as soon as you dig a mm under the surface - nobody has tested it, it’s got a critical bug, it is overblown and there are issues with no response. No docs, no info, no nothing. For the biggest thing in technology in history, it is awfully hollow. I don’t mean it in a condescending way, in fact community is enthusiastic and very helpful, it’s just that it doesn’t live up to what most would expect.
A caveat I need to mention is I have not used it for coding - I have an irrational fear and resistance towards it, being a programmer. I just won’t touch it, even if it means the end of my career. I’m trying to be grown-up about it, but so far, I dont want to use it, for good and bad reasons.
Yes, I got a Strix Halo machine before the RAM price hike and use it to run all my ML stuff on it.
Currently using llama-swap with llama.cpp/ComfyUI and opencode/Open WebUI as frontend.
I’m running Qwen3.6-27b, Voxtral Mini 4b, Piper and Qwen Image. Also, some embedding and reranking models.
I use them for:
- Tagging and classification of my documents in Paperless
- Home Assistant (voice assistant)
- Translations (both text and image)
- Transcriptions
- Some light coding and debugging
- Avatar/Backdrop generation for DnD sessions
What sort of tok/s are you getting on the strix?
About 200 t/s prompt processing and 10-20 t/s with MTP.
Greatly depends on the task, predictable things like code generates at 18-20 t/s. Creative writing more like 10-17 t/s.
Damn - I thought strix would do a bit better than that, for how much it costs.
Given the 27b is a dense model, I think the numbers are quite ok. Curious about the quant tho.
The cool thing about the strix is its large unified memory, but it lacks memory bandwith for compute intensive workloads. Something like Qwen3.5-122b MoE with only like 12b active parameters might run at twice the speed if it fits the configuration.
Curious about the quant tho.
Q8 from unsloth.
Something like Qwen3.5-122b
My go to model for knowledge. Definitely much faster at Q5 but it lacks the tool calling quality of the Qwen3.6 models. Really hoping we see a Qwen3.6-122b soon…
In case you missed the Ornith 1.0 release (Qwen and Gemma RL finetunes for agentic / coding workloads), they look interesting to bridge the gap until we see larger 3.6 models or a 3.7 release. I didn’t test them yet but according to benchmarks, the 35b MoE seems to be more or less on par with Qwen3.6 27b dense, while ofc a lot faster.
Yeah. Though I think theres a new strix out soon (Medusa? Gorgon? Something like that).
Its a bit like my P40. On paper, it has 24GB. But that 24gb is capped at 400GB/s and the ai compute is what…Pascal era?
AI = Good, fast, cheap - pick 2
Well compared to the strix, 400GB/s is not that bad, I think with fast system RAM and expert offloading you could squeeze quite something out of it when running stuff in the 100b-a10b regions.
Your bigger problem is going to be future software support.
No. I still have no use for it and everything I use is automated without at a far lower footprint.
I set up ollama on our thinkstation in the lab and I use it for looking up documentation, generating readmes, searching papers, and sometimes coding when I know what to do but don’t feel it is worth it to spend time on it myself. So basically the chat with web search.
Which models did you find particularly useful for those tasks?
Gemma 4, gpt oss, and nemotron. Currently I’ve been sticking with Gemma more, the 31 billion parameters one.
I currently run Qwen3.6-27b on llama.cpp and use it via openwebui. Mostly, I use it for web research via tavily, to a lesser extent for coding and interactively learning about things that are new to me but common in training data (such as basic math or ML concepts).
Yup, ollama, various models. I initially downloaded it because I, along with thousands of other people, wanted to see what would happen if I made models debate with each other after RAGging them with various books (The Prince, The Art of War, The complete works of Shakespeare, etc.).
The results were uninteresting and I abandoned the project pretty quickly. I’ll sometimes use them for code analysis but they’re too slow on my rig to be really useful.
wanted to see what would happen if I made models debate
LOL I kind of do that…sort of. I’ll ask several AI the very same question to see what they spit out.
You’ll like this then
Well I’ll be damned. Of course the law of large numbers dictates someone, somewhere has the same thought.
One of the projects I started and never got to a satisfactory end state was basically that, plus a judging round. Every model would respond to the same prompt, then every model would evaluate every other model’s response for accuracy and completeness. Then the results would get logged to a spreadsheet.
It’s simple enough, but for N models it requires N + N^2 model calls so it takes forever to run any decent dataset on consumer hardware. If I had the resources and a way to run it that didn’t fry the planet, I think it would be a cool running set of comparative benchmarks. IDK if it’d be useful at all but I’m still interested to see the data.
Every model would respond to the same prompt, then every model would evaluate every other model’s response for accuracy and completeness
If I understand correctly I sorta kinda do that. I’ll copy and paste one AI’s response into another and prompt something like 'Validate AI response: and paste it in. HAHA I thought I was being tricky but you’re already on it.
I think it’s tricky. It’s kind of like adding LLMs like vectors, and hopefully the effect can soften or at least reveal the shortcomings of individual models. Is it a good idea? I don’t know, I think there are good reasons to think it’s a waste of time and resources. I certainly think I’d need a better explanation of what use it would be before I spent more time building it. But I still think about what use it would be from time to time; I haven’t decided that it’s a bad idea yet.
at least reveal the shortcomings of individual models. Is it a good idea? I don’t know,
I mean I do it, in my rudimentary way, to check for some semblance of consistency. I’m unclear why you think that not a good idea?
P.S. This is a hypothesis, I haven’t even designed the test for it, much less run it. What follow are my suppositions.
I think whether or not it’s a good idea depends on how similar all the models are. I don’t have a rigorous definition of “similar” but things like similar training data, similar design methodologies, similar QA processes would all contribute. Theoretically (I think), if they’re all dissimilar, they should each catch errors the others miss. However, the more similar they are, the more likely they have the same biases and weak spots, and your error rate from a response + verification may be the same or even higher than the error rate for just the original prompt, and you’d be unlikely to detect those errors using just two similar models. It can instill false confidence in the results because you’re doing something that should in theory increase the validity of the data, but in practice might make no difference or even make the quality of responses worse.
Did you use OWUIs native “call simultaneous models to answer” feature for that or one of the AI debate harnesses?
Nothing so fancy. I just made a little python script to prompt the first model, wait for a response, then prompt the next model with the initial prompt + the response, and so on. It was very hacky and slow.
Ah - I thought you might have used something like this
Oh neat. Yeah, if something like that had existed (and I’d been aware of it) I probably would have used it instead of building my own shoestring version.
Ollama with gemma 4 for LLM stuff, coding brainstorming, etc.
Comfy ui with z-image or stable diffusion for images.
How is Gemma compared to others?
I tried but I only have 16g of ram and it wouldn’t complete a thought alas
If I wanted AI for some reason, it’d be self-host or nothing.
Running decencored Qwen3.6-27b and a 9b Gemma for RAG and scrapes on Ollama with a mostly vibe coded discord bot. Just got it to run tools and scrape and post news on a schedule. The first model I can run locally that’s smart enough to be useful. May give Jan a try for the back end after reading that other guys rant.
Mostly use it for stupid questions I could have googled and to brag to friends.
I ran through lmstudio because it really eazy, I ran some kind of qwen 3.6 27b imatrix neo code DI, it is the best local model for coding I tried, I think it can be better than some cloud model
Jup. Ollama and OpenWebUI is a great stack to tinker with some LLM models. They’re kinda useful for aggregating large datasets, translations, frontend development and gathering relevant sources for me to read into. Also, Qwen has been amazing in understanding frameworks without documentation and writing one for me. I had to use some self-developed PHP framework for a task once and without qwen, I would’ve taken probably two more weeks to get the task done.
MiniCPM has also been REALLY good at image detection, describing it as accurately as possible, feeding it into qwen who then searches what the object could be and returning the result. I always liked google lense and that stack gave me a TEMU-Version of google lense that isn’t quite as reliable, but definitely very useful.
This sounds like an excellent use case and I don’t know why you were downvoted.
Because AI bad I guess. I’m not expecting sane behavior on lemmy when it comes to AI or capitalism lmao
Probably. I wish Lemmy would remove up / down votes entirely. I might ask our lovely new mod if that’s possible.
The like / dislike button has been a curse since invented.
https://www.theguardian.com/technology/2017/oct/05/smartphone-addiction-silicon-valley-dystopia
EDIT: ah - I can hide em on my end. Not as good but it will do
