Was wondering if I was to buy cheapest hardware (eg PC) to run for personal use at reasonable speed llama 2 70b what would that hardware be? Any experience or recommendations?
Anything with 64GB of memory will run a quantized 70B model. What else you need depends on what is acceptable speed for you. With a decent CPU but without any GPU assistance, expect output on the order of 1 token per second, and excruciatingly slow prompt ingestion. Any decent Nvidia GPU will dramatically speed up ingestion, but for fast generation, you need 48GB VRAM to fit the entire model. That means 2x RTX 3090 or better. That should generate faster than you can read.<p>Edit: the above is about PC. Macs are much faster at CPU generation, but not nearly as fast as big GPUs, and their ingestion is still slow.
I built a DIY PC with used GPUs (2x RTX 3090) for around 2300€ earlier this year. You can probably do it for slightly less now (i also added 128GB RAM and NVLink). You can generate text with >10 tok/s with that setup.
Make sure to get a PSU with more than 1000W.
Air cooling is a challenge, but it's possible.<p>Recommended reading: Tim Dettmer's guide <a href="https://timdettmers.com/2023/01/30/which-gpu-for-deep-learning/" rel="nofollow noreferrer">https://timdettmers.com/2023/01/30/which-gpu-for-deep-learni...</a>
If you have a lot of money (but not H100/A100 money), get 4090s as they're currently the best bang for your buck on the CUDA side (according to George Hotz). If broke, get multiple second hand 3090s. <a href="https://timdettmers.com/2023/01/30/which-gpu-for-deep-learning/" rel="nofollow noreferrer">https://timdettmers.com/2023/01/30/which-gpu-for-deep-learni...</a>. If unwilling to spend any money at all and just want to play around with llama70b, look into petals <a href="https://github.com/bigscience-workshop/petals">https://github.com/bigscience-workshop/petals</a>
We bought an A6000 48GB ( as mentioned by someone else ) and it’s works great for $3800. The power requirements are modest as well compared to consumer GPU’s. We looked at the ADA version but even used they are a lot more and your buying speed not usability. I would rather buy another A6000 and have 96GB of ram to fine tune with. That’s just me though and everyone needs to rank their needs against what they can afford.
A 192gb Mac Studio should be able to run an unquantized 70B and I think would cost less than running a multi gpu setup made up of nvidia cards. I haven’t actually done the math, though.
If you factor in electricity costs over a certain time period it might make the Mac even cheaper!
The only info I can provide is the table I've seen on: <a href="https://github.com/jmorganca/ollama">https://github.com/jmorganca/ollama</a> where it states one needs "32 GB to run the 13B models." I would assume you may need a GPU for this.<p>Related, could someone please point me in the right direction on how to run Wizard Vicuna Uncensored or Llama2 13B locally in Linux? I've been searching for a guide and have not found what I need for a beginner like myself. In the Github I referenced the download is only for Mac at the time. I have a Macbook Pro M1 I can use though it's running Debian.<p>Thank you.
I've been able to run it fine using llama.cpp on my 2019 iMac with 128GB of RAM. It's not super fast, but it works fine for "send it a prompt, look at the reply a few minutes later", and all it cost me was a few extra sticks of RAM.
You can run on cpu and regular ram, but gpu is quite a bit faster.<p>You need about a gig of RAM/nvram per billion parameters (plus some headroom for a context window). Lower precision doesn’t really affect quality.<p>When Ethereum flipped from proof of work to proof of stake, a lot of used high-end cards hit the market.<p>4 of them in a cheap server would do the trick. Would be a great business model for some cheap colo to stand up a crap-ton of those and rent while servers to everyone here.<p>In the meantime if you’re interested in a cheap server as described above, post in this thread.
I don't think it is the cheapest, but the tiny box is an option:<p><a href="https://tinygrad.org/" rel="nofollow noreferrer">https://tinygrad.org/</a>
I feel as if the cheapest way of running these kinds of models would be to have the whole cache/memory take space on the hard drive rather than the RAM. Then, you could just use CPU power instead of splurging out thousands for RAM & a GPU with enough VRAM.<p>It might or might not be reasonable speeds, but I would reason that it could avoid "sunk cost irony"; if you decide, that any point, Chat-GPT would have sufficed in your task. It's rare, but it can happen.<p>If you want to take this silly logic further, you can theoretically run any sized model on any computer. You could even attempt this dumb idea on a computer running Windows 95. I don't care how long it would take; if it takes seven and a half million years for 42 tokens, I would still call it a success!
If it's only for a short time, use a price calculator to decide if it's worth renting GPUs on a cloud provider. You can get immediate temporary access for far more computing power than you can ever hope to buy outright.