brucethemoose

joined 1 year ago
[–] brucethemoose@lemmy.world 13 points 2 days ago* (last edited 2 days ago) (3 children)

The implication is walking away all US military support, I believe.

[–] brucethemoose@lemmy.world 6 points 2 days ago (1 children)
[–] brucethemoose@lemmy.world 51 points 2 days ago* (last edited 2 days ago) (8 children)

I’m hoping Arc survives all this?

I know they want to focus, but no one’s going to want their future SoCs if the GPU part sucks or is nonexistent. Heck, it’s important for servers, eventually.

Battlemage is good!

[–] brucethemoose@lemmy.world 0 points 4 days ago

It was selectively given to institutions and "major" celebrities before that.

Selling them dilutes any meaning of "verified" because any joe can just pay for extra engagement. It's a perverse incentive, as the people most interest in grabbing attention buy it and get amplified.

It really has little to do with Musk.

[–] brucethemoose@lemmy.world 5 points 4 days ago* (last edited 4 days ago) (1 children)

the whole concept is stupid.

+1

Being that algorithmic just makes any Twitter-like design too easy to abuse.

Again, Lemmy (and Reddit) is far from perfect, but fundamentally, grouping posts and feeds by niche is way better. It incentivizes little communities that are concerned about their own health, while users have zero control over that shouting into the Twitter maw.

[–] brucethemoose@lemmy.world 7 points 4 days ago* (last edited 4 days ago)

Not sure where you're going with that, but it's a perverse incentive, just like the engagement algorithm.

Elon is a problem because he can literally force himself into everyone's feeds, but also because he always posts polarizing/enraging things these days.

Healthy social media design/UI is all about incentivizing good, healthy communities and posts. Lemmy is not perfect, but simply not designing for engagement/profit because Lemmy is "self hosted" instead of commercial is massive.

[–] brucethemoose@lemmy.world 3 points 1 week ago* (last edited 1 week ago)

onlyfans is successful because you can pay for personalized actions and one can build something appearing like a relationship with the porn star. Before one would just passively watch them.

It’s still so one-way though.

I guess what’s remarkable is the collective unawareness of how artificial those relationships are. I get it, I know IRL loneliness and went down the engagement rabbit hole too.

But still, it’s remarkable. I still see articles from professional journalists, all the time, wondering what’s happening to relationships or democracy or whatever then end their post with something akin to a “like and subscribe!” and a busy Twitter profile.

[–] brucethemoose@lemmy.world 8 points 1 week ago (3 children)

It’s great for pushing enraging content though. I feel like the content itself has gravely affected the collective psyche, and so have influencer endorsements (which companies pay big bucks for), even if the explicit ads have not.

[–] brucethemoose@lemmy.world 7 points 1 week ago* (last edited 1 week ago) (1 children)

Look at her social media posting history. Look at her published book:

BRAVE Books and Ashley St. Clair partnered to write "Elephants Are Not Birds," a Christian, Conservative children's book that tackles the topic of gender identity. In the book, children will learn that boys are not girls, and Elephants Are Not Birds.

I retract/regret what I said… I didn’t mean to victim blame. This sucks for her and she doesn’t deserve one bit of it, nor does her abuser deserve any sympathy. She deserves empathy.

But she also appears to have made a living peddling (IMO) hate far and wide. I am struggling to express/process how I feel about that… it’s a tragic irony, I guess? And frustration at that whole ecosystem she was in (which Musk is at the very tippy top of).

[–] brucethemoose@lemmy.world 1 points 1 week ago* (last edited 1 week ago) (1 children)

Watch this video, ignore the clickbait sounding title:

https://youtu.be/Ufmu1WD2TSk

It completely changed my view on that.

Basically, without high birth rates, countries are totally screwed. Immigration (which skews young, from high birth rate countries), has softened that issue for the US, hence you don’t hear about it as much here. One can wave their hands and say "elder care and the economy will be automated in the future," but that’s wishful thinking if you ask me.

Figuring out how to more efficiently house/care for a glut of humans farther in the future is way more practical. Honestly we’re ridiculously inefficient now; there’s a lot of low hanging fruit to pick. And we can use much higher technology to address that.

[–] brucethemoose@lemmy.world 2 points 1 week ago* (last edited 1 week ago) (6 children)

Yeah, at the very least she needs security and lawyers. 20K is little enough to endanger her life.

…On the other hand, she's an influencer farming attention. And not in a good way. That cuts a lot of my empathy short.

[–] brucethemoose@lemmy.world 18 points 1 week ago (1 children)

It’s really not funny TBH, it’s all tied to his views on eugenics (or whatever euphemism is used) and frankly horrific visions of the future he shares with a few lower profile billionaires.

 

I see a lot of talk of Ollama here, which I personally don't like because:

  • The quantizations they use tend to be suboptimal

  • It abstracts away llama.cpp in a way that, frankly, leaves a lot of performance and quality on the table.

  • It abstracts away things that you should really know for hosting LLMs.

  • I don't like some things about the devs. I won't rant, but I especially don't like the hint they're cooking up something commercial.

So, here's a quick guide to get away from Ollama.

  • First step is to pick your OS. Windows is fine, but if setting up something new, linux is best. I favor CachyOS in particular, for its great python performance. If you use Windows, be sure to enable hardware accelerated scheduling and disable shared memory.

  • Ensure the latest version of CUDA (or ROCm, if using AMD) is installed. Linux is great for this, as many distros package them for you.

  • Install Python 3.11.x, 3.12.x, or at least whatever your distro supports, and git. If on linux, also install your distro's "build tools" package.

Now for actually installing the runtime. There are a great number of inference engines supporting different quantizations, forgive the Reddit link but see: https://old.reddit.com/r/LocalLLaMA/comments/1fg3jgr/a_large_table_of_inference_engines_and_supported/

As far as I am concerned, 3 matter to "home" hosters on consumer GPUs:

  • Exllama (and by extension TabbyAPI), as a very fast, very memory efficient "GPU only" runtime, supports AMD via ROCM and Nvidia via CUDA: https://github.com/theroyallab/tabbyAPI

  • Aphrodite Engine. While not strictly as vram efficient, its much faster with parallel API calls, reasonably efficient at very short context, and supports just about every quantization under the sun and more exotic models than exllama. AMD/Nvidia only: https://github.com/PygmalionAI/Aphrodite-engine

  • This fork of kobold.cpp, which supports more fine grained kv cache quantization (we will get to that). It supports CPU offloading and I think Apple Metal: https://github.com/Nexesenex/croco.cpp

Now, there are also reasons I don't like llama.cpp, but one of the big ones is that sometimes its model implementations have... quality degrading issues, or odd bugs. Hence I would generally recommend TabbyAPI if you have enough vram to avoid offloading to CPU, and can figure out how to set it up. So:

This can go wrong, if anyone gets stuck I can help with that.

  • Next, figure out how much VRAM you have.

  • Figure out how much "context" you want, aka how much text the llm can ingest. If a models has a context length of, say, "8K" that means it can support 8K tokens as input, or less than 8K words. Not all tokenizers are the same, some like Qwen 2.5's can fit nearly a word per token, while others are more in the ballpark of half a work per token or less.

  • Keep in mind that the actual context length of many models is an outright lie, see: https://github.com/hsiehjackson/RULER

  • Exllama has a feature called "kv cache quantization" that can dramatically shrink the VRAM the "context" of an LLM takes up. Unlike llama.cpp, it's Q4 cache is basically lossless, and on a model like Command-R, an 80K+ context can take up less than 4GB! Its essential to enable Q4 or Q6 cache to squeeze in as much LLM as you can into your GPU.

  • With that in mind, you can search huggingface for your desired model. Since we are using tabbyAPI, we want to search for "exl2" quantizations: https://huggingface.co/models?sort=modified&search=exl2

  • There are all sorts of finetunes... and a lot of straight-up garbage. But I will post some general recommendations based on total vram:

  • 4GB: A very small quantization of Qwen 2.5 7B. Or maybe Llama 3B.

  • 6GB: IMO llama 3.1 8B is best here. There are many finetunes of this depending on what you want (horny chat, tool usage, math, whatever). For coding, I would recommend Qwen 7B coder instead: https://huggingface.co/models?sort=trending&search=qwen+7b+exl2

  • 8GB-12GB Qwen 2.5 14B is king! Unlike it's 7B counterpart, I find the 14B version of the model incredible for its size, and it will squeeze into this vram pool (albeit with very short context/tight quantization for the 8GB cards). I would recommend trying Arcee's new distillation in particular: https://huggingface.co/bartowski/SuperNova-Medius-exl2

  • 16GB: Mistral 22B, Mistral Coder 22B, and very tight quantizations of Qwen 2.5 34B are possible. Honorable mention goes to InternLM 2.5 20B, which is alright even at 128K context.

  • 20GB-24GB: Command-R 2024 35B is excellent for "in context" work, like asking questions about long documents, continuing long stories, anything involving working "with" the text you feed to an LLM rather than pulling from it's internal knowledge pool. It's also quite goot at longer contexts, out to 64K-80K more-or-less, all of which fits in 24GB. Otherwise, stick to Qwen 2.5 34B, which still has a very respectable 32K native context, and a rather mediocre 64K "extended" context via YaRN: https://huggingface.co/DrNicefellow/Qwen2.5-32B-Instruct-4.25bpw-exl2

  • 32GB, same as 24GB, just with a higher bpw quantization. But this is also the threshold were lower bpw quantizations of Qwen 2.5 72B (at short context) start to make sense.

  • 48GB: Llama 3.1 70B (for longer context) or Qwen 2.5 72B (for 32K context or less)

Again, browse huggingface and pick an exl2 quantization that will cleanly fill your vram pool + the amount of context you want to specify in TabbyAPI. Many quantizers such as bartowski will list how much space they take up, but you can also just look at the available filesize.

  • Now... you have to download the model. Bartowski has instructions here, but I prefer to use this nifty standalone tool instead: https://github.com/bodaay/HuggingFaceModelDownloader

  • Put it in your TabbyAPI models folder, and follow the documentation on the wiki.

  • There are a lot of options. Some to keep in mind are chunk_size (higher than 2048 will process long contexts faster but take up lots of vram, less will save a little vram), cache_mode (use Q4 for long context, Q6/Q8 for short context if you have room), max_seq_len (this is your context length), tensor_parallel (for faster inference with 2 identical GPUs), and max_batch_size (parallel processing if you have multiple user hitting the tabbyAPI server, but more vram usage)

  • Now... pick your frontend. The tabbyAPI wiki has a good compliation of community projects, but Open Web UI is very popular right now: https://github.com/open-webui/open-webui I personally use exui: https://github.com/turboderp/exui

  • And be careful with your sampling settings when using LLMs. Different models behave differently, but one of the most common mistakes people make is using "old" sampling parameters for new models. In general, keep temperature very low (<0.1, or even zero) and rep penalty low (1.01?) unless you need long, creative responses. If available in your UI, enable DRY sampling to tamp down repition without "dumbing down" the model with too much temperature or repitition penalty. Always use a MinP of 0.05 or higher and disable other samplers. This is especially important for Chinese models like Qwen, as MinP cuts out "wrong language" answers from the response.

  • Now, once this is all setup and running, I'd recommend throttling your GPU, as it simply doesn't need its full core speed to maximize its inference speed while generating. For my 3090, I use something like sudo nvidia-smi -pl 290, which throttles it down from 420W to 290W.

Sorry for the wall of text! I can keep going, discussing kobold.cpp/llama.cpp, Aphrodite, exotic quantization and other niches like that if anyone is interested.

view more: next ›