this post was submitted on 24 Jul 2024

1 points (100.0% liked)

Technology

58480 readers

3893 users here now

This is a most excellent place for technology news and articles.

Our Rules

Follow the lemmy.world rules.
Only tech related content.
Be excellent to each another!
Mod approved content bots can post up to 10 articles per day.
Threads asking for personal tech support may be deleted.
Politics threads may be removed.
No memes allowed as posts, OK to post as comments.
Only approved bots from the list below, to ask if your bot can be added please contact us.
Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago

MODERATORS

The first GPT-4-class AI model anyone can download has arrived: Llama 405B (arstechnica.com)

submitted 2 months ago by Wilshire@lemmy.world to c/technology@lemmy.world

57 comments fedilink hide all child comments

you are viewing a single comment's thread
view the rest of the comments

[–] admin@lemmy.my-box.dev 0 points 2 months ago (4 children)

Technically correct (tm)

Before you get your hopes up: Anyone can download it, but very few will be able to actually run it.

[–] coffee_with_cream@sh.itjust.works 0 points 2 months ago* (last edited 2 months ago) (1 children)

This would probably run on a a6000 right?

Edit: nope I think I'm off by an order of magnitude

[–] 5redie8@sh.itjust.works 0 points 2 months ago

"an order of magnitude" still feels like an understatement LOL

My 35b models come out at like Morse code speed on my 7800XT, but at least it does work?

[–] bitfucker@programming.dev 0 points 2 months ago

So does OSM data. Everyone can download the whole earth but to serve it and provide routing/path planning at scale takes a whole other skill and resources. It's a good thing that they are willing to open source their model in the first place.

[–] chiisana@lemmy.chiisana.net 0 points 2 months ago (3 children)

What’s the resources requirements for the 405B model? I did some digging but couldn’t find any documentation during my cursory search.

[–] modeler@lemmy.world 0 points 2 months ago* (last edited 2 months ago) (6 children)

Typically you need about 1GB graphics RAM for each billion parameters (i.e. one byte per parameter). This is a 405B parameter model. Ouch.

Edit: you can try quantizing it. This reduces the amount of memory required per parameter to 4 bits, 2 bits or even 1 bit. As you reduce the size, the performance of the model can suffer. So in the extreme case you might be able to run this in under 64GB of graphics RAM.

[–] cheddar@programming.dev 0 points 2 months ago

Typically you need about 1GB graphics RAM for each billion parameters (i.e. one byte per parameter). This is a 405B parameter model.

[–] 1984@lemmy.today 0 points 2 months ago

Can you run this in a distributed manner, like with kubernetes and lots of smaller machines?

[–] obbeel@lemmy.eco.br 0 points 2 months ago

According to huggingface, you can run a 34B model using 22.4GBs of RAM max. That's a RTX 3090 Ti.

[–] TipRing@lemmy.world 0 points 2 months ago

When the 8 bit quants hit, you could probably lease a 128GB system on runpod.

[–] Siegfried@lemmy.world 0 points 2 months ago* (last edited 2 months ago)

At work we habe a small cluster totalling around 4TB of RAM

It has 4 cooling units, a m3 of PSUs and it must take something like 30 m2 of space

[–] Deceptichum@quokk.au 0 points 2 months ago (2 children)

Or you could run it via cpu and ram at a much slower rate.

[–] errer@lemmy.world 0 points 2 months ago (1 children)

Yeah uh let me just put in my 512GB ram stick…

[–] Deceptichum@quokk.au 0 points 2 months ago

Samsung do make them.

Goodluck finding 512gb of VRAM.

[–] chiisana@lemmy.chiisana.net 0 points 2 months ago

Finally! My dumb dumb 1TB ram server (4x E5-4640 + 32x32GB DDR3 ECC) can shine.

[–] sunzu@kbin.run 0 points 2 months ago (2 children)

405b ain't running local unless you got a proepr set up is enterpise grade lol

I think 70b is possible but I haven't find anyone confirming it yet

Also would like to know specs on whoever did it

[–] raldone01@lemmy.world 0 points 2 months ago (1 children)

I regularly run llama3 70b unqantesized on two P40s and CPU at like 7tokens/s. It's usable but not very fast.

[–] sunzu@kbin.run 0 points 2 months ago (2 children)

so there is no way a 24gb and 64gb can run thing?

[–] raldone01@lemmy.world 0 points 2 months ago* (last edited 2 months ago) (1 children)

My specs because you asked:

CPU: Intel(R) Xeon(R) E5-2699 v3 (72) @ 3.60 GHz
GPU 1: NVIDIA Tesla P40 [Discrete]
GPU 2: NVIDIA Tesla P40 [Discrete]
GPU 3: Matrox Electronics Systems Ltd. MGA G200EH
Memory: 66.75 GiB / 251.75 GiB (27%)
Swap: 75.50 MiB / 40.00 GiB (0%)

[–] sunzu@kbin.run 0 points 2 months ago (1 children)

ok this is a server. 48gb cards and 67gb ram? for model alone?

[–] raldone01@lemmy.world 0 points 2 months ago

Each card has 24GB so 48GB vram total. I use ollama it fills whatever vrams is available on both cards and runs the rest on the CPU cores.

[–] raldone01@lemmy.world 0 points 2 months ago (1 children)

What are you asking exactly?

What do you want to run? I assume you have a 24GB GPU and 64GB host RAM?

[–] sunzu@kbin.run 0 points 2 months ago (1 children)

correct. and how ram speed work in this tbh

[–] raldone01@lemmy.world 0 points 2 months ago

My memory sticks are all DDR4 with 32GB@2133MT/s.

[–] Voyajer@lemmy.world 0 points 2 months ago* (last edited 2 months ago) (2 children)

I've run quantized 70B models on CPU with 32 gigs but it is very slow

[–] bizarroland@fedia.io 0 points 2 months ago (1 children)

I have a home server with 140 gigs of RAM, it was surprisingly cheap. It's an HP z6 with the 6146 gold xeon processor.

I found a seller who was selling it with a low spec silver and 16 gigs of RAM for like 250 bucks.

Found the processor upgrade for about $120 and spend another $150 on 128gb of second-hand ECC ddr4.

I think the total cost was something like $700 after throwing a couple of 8 TB hard drives in.

I've also placed a Nvidia 4070 in it, which I got doing some horse trading.

How close am I on the specs to being able to run the 70b version?

[–] BaroqueInMind@lemmy.one 0 points 2 months ago (1 children)

What's the bus speed of the RAM? You might run it just find but still bottlenecked there.

[–] bizarroland@fedia.io 0 points 2 months ago (1 children)

It's clocked at ddr4 2666

[–] BaroqueInMind@lemmy.one 0 points 2 months ago* (last edited 2 months ago)

With 144Gb of total RAM, you should be able to run any CPU intensive software.

The LLMs use GPU vRAM though, so it doesn't matter how much system RAM you have, since GPU vRAM is what the xformers and tensor scripts prioritize and have been ultimately optimized to use over CPU and RAM.

[–] sunzu@kbin.run 0 points 2 months ago

I gonna add some RAM with hope I can split original 70b between GPU and RAM. 8b is great what it is as is

Looks like it should be possible, not sure how much performance hit offloading to RAM will do. Fafo

[–] Blaster_M@lemmy.world 0 points 2 months ago

As a general rule of thumb, you need about 1 GB per 1B parameters, so you're looking at about 405 GB for the full size of the model.

Quantization can compress it down to 1/2 or 1/4 that, but "makes it stupider" as a result.

[–] LavenderDay3544@lemmy.world 0 points 2 months ago (1 children)

When RTX 9090 Ti comes anyone who can afford it will be able to run it.

[–] Contravariant@lemmy.world 0 points 2 months ago

That doesn't sound like much of a change from the situation right now.