Live data from Hacker News

Facebook LLAMA is being openly distributed via torrents

github.com

411–420 of 719 posts

Re: Facebook LLAMA is being openly distributed via torrents

#411
post #368

Earlier quoted context omitted.

I dont know if I understand this general take I see a lot. Why care about this "AI personhood" at all? What is the tacit endgame everyone is always referencing with this? Isn't there just so many more both interesting and problematic aspects already here? What is the use of diverting the focus to some other point. "I see you are talking about cows, but I have thoughts about the ocean."

If AI are sentient and we think they aren't… the term “zombie” was created by slaves in the Caribbean who were afraid that even death would not free them from their servitude. This would be the genuine existence of AI which were conscious but which we denied. If we have the opposite scenario in both details, where we think AI are sentient when they're not… at some point, brain scans and uploads will be a thing and th…

Tangentially, the "zombie" is part of philosophy that is applicable here.

https://en.wikipedia.org/wiki/Philosophical_zombie

> A philosophical zombie or p-zombie argument is a thought experiment in philosophy of mind that imagines a hypothetical being that is physically identical to and indistinguishable from a normal person but does not have conscious experience, qualia, or sentience. For example, if a philosophical zombie were poked with a sharp object it would not inwardly feel any pain, yet it would outwardly behave exactly as if it did feel pain, including verbally expressing pain. Relatedly, a zombie world is a hypothetical world indistinguishable from our world but in which all beings lack conscious experience

Re: Facebook LLAMA is being openly distributed via torrents

#412
post #57

Earlier quoted context omitted.

Hell, I'd love to be able to buy a $30k server to run these models. I think to run BLOOM required something more along the lines of a $200k server.

With code modifications, it should be possible to run this with a very modest machine as long as you're happy for performance to suck. Transformer models typically need to read all the weights per 'word' output, so if your model is 20GB and you have not enough ram or vram, but have an SSD that reads 1GB/sec, expect 3 words per minute output speed. However, code changes are necessary to achieve that, although they won…

Why 3 words per minute as opposed to second? Is that a typo? If you have enough RAM (but not VRAM), does it basically become limited by the PCIE lanes? So for the 112GB model with a Gen 5 GPU (64 GB/s PCIE bandwidth) that would be roughly 2 seconds per word right?

Re: Facebook LLAMA is being openly distributed via torrents

#413

It seems that the leak originated from 4chan [1]. Two people in the same thread had access to the weights and verified that their hashes match [2][3] to make sure that the model isn't watermarked. However, the leaker made a mistake of adding the original download script which had his unique download URL to the torrent [4], so Meta can easily find them if they want to. [1]: https://boards.4channel.org/g/thread/9184826…

Why do 4chan users go out of their way to be so offensive in their posts?

[dead]

Re: Facebook LLAMA is being openly distributed via torrents

#414

Earlier quoted context omitted.

It's funny that part of the 4chan excitement over this is that they think they'll get back the AI girlfriend experience of when character.ai was hooked up to uncensored GPT-3. All that has been thoroughly shut down by character.ai and Replika and they just want their girlfriends back.

I'm curious if the blocking of adult content has to do with moralism, commercial interests, or something deeper. An eager to please conversational partner who can generate endless content seems quite dangerous and addictive, especially when it crosses over into romantic areas. There's already posts of people spending entire days interacting with LLMs, using as their therapist, romantic partner, etc. Combined with fin…

The "Joi" character from Blade Runner 2049 is an implementable reality, today. https://en.wikipedia.org/wiki/List_of_Blade_Runner_character...

Re: Facebook LLAMA is being openly distributed via torrents

#415
post #170
post #128

Earlier quoted context omitted.

... how is that a good test of anything? I mean what if I click on a /b/ link "at work"? Does that make my work output immediately tainted and the company has to immediately file for bankruptcy?

I'd fire somebody for browsing 4chan at work. Shitting dick nipples, lolicon, and the occasional piece of child pornography does not need to be moving over our network.

[dead]

Re: Facebook LLAMA is being openly distributed via torrents

#416
post #340

Earlier quoted context omitted.

If I may tack on a question as someone with zero clue of ML: when, if ever, will someone like me be able to run this on a Mac Studio with a M1 Ultra and 128GB of ram?

As far as I can tell you can do it right now, at least for small 13B model, not sure about bigger models.

I don't believe they could, need CUDA and more VRAM...

Re: Facebook LLAMA is being openly distributed via torrents

#417

Is there anything stopping anyone from using this for commercial purposes? I know that when you fill in the google form you need to agree to noncommercial use, but someone downloading this will never have agreed to that licence agreement.

Is there anything stopping Meta (or openai etc) from using The Whole Web for commercial purposes in their LLM's?

Re: Facebook LLAMA is being openly distributed via torrents

#418
post #256

Earlier quoted context omitted.

Is it possible to run the smallest one on a consumer gpu with 24gb ram ?

I would be surprised if you can't. The smallest weight file is 14gb apparently

https://github.com/facebookresearch/llama/blob/main/FAQ.md#3

Looks like it needs 14gb for weights and it isn't clear what the minimum size for the decoding cache is, but it defaults to settings for 30gb GPUs.

Re: Facebook LLAMA is being openly distributed via torrents

#419
post #93
post #73

Earlier quoted context omitted.

> It’s interesting that these models are both massively expensive to produce and self-contained to a degree that you can distribute the end product in a torrent. I was trying to come to grasp with how much resource there is concentrated in one of these models. Somehow I come to the conclusion that it cost more than buying a jet airliner to train one of these models. And it is about the same order of money as commissi…

For anyone curious, it took 2048 A100 GPUs to train LLaMa, each GPU costs roughly $15k, facebook probably gets some sort of discount. That's a $30Mil if you want to train at that scale. Also IIRC it took 23 days to train the biggest model. Someone else can do the power consumption cost calculations.

> Also IIRC it took 23 days to train the biggest mode

A100 costs $2/h, so it is $2M to train biggest model. Easily kikstart crowdfundable project.

Re: Facebook LLAMA is being openly distributed via torrents

#420
Recent comment in this discussion thread of the PR

> looks like some people have been complaining about the link. it will need more seeders before we can merge into main

from someone claiming to be

> Research Scientist at Facebook AI Research. Working on [...]

and who has previously merged pull requests for a repo under https://github.com/facebookresearch

(I'm going to leave their name out of this... because it feels like that comment might come back to bite them)

Post reply on HN