Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

twitter.com

Post reply on HN
Print view View original

101–110 of 143 posts

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#101

Post by EVa5I7bHFq9mnYK »

dragonwriter wrote: post #52
mdaniel wrote: post #44

Earlier quoted context omitted.

To add to this excellent reply, I'll also point out that the reason folks want the weights is that they are the result of a massive search operation, akin to finding the right temperature to bake a cake from all possible floats. It takes a lot of wall clock time, and a lot of GPU energy, and a lot of input examples and counter-examples to find the "right" numbers. Thus, it really is better -- all things being equal -…

> a massive search operation, akin to finding the right temperature to bake a cake from all possible floats ...for each of 13 billion (for a model with that many parameters) different cakes, except that they aren’t like cakes because the “best" temperature for each depends on the actual temperatures chosen for the others.

It's 2^(16*13,000,000,000) different cakes.

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#102

Post by ithkuil »

oh_sigh wrote: post #91
stcredzero wrote: post #71

Earlier quoted context omitted.

For many humans, some "hobbies" involve "projects" which may involve seemingly infinite degrees of procrastination. (This certainly applies to me!)

You're not wrong - but for perspective this is equivalent to a 90 year old picking up a hobby 5 hours ago.

Is it though? It could be a child picking up a hobby after being old enough to appreciate the hobby. There is so much more time left in the universe before heat death, so the 90y metaphor doesn't really describe the current point in time

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#103

Post by muyuu »

gigel82 wrote: post #95

One of my favorite queries to assess an LLM is Is it legal for a man to marry his widow's sister? Vicuna-13B: The legality of a man marrying his widow's sister (widow marriage) varies depending on the country and cultural context. In some societies, widow marriage was traditionally practiced or tolerated. GPT-4: It is not possible for a man to marry his widow's sister, as the term "widow" refers to a deceased man's w…

would still be possibly legal on the basis that if it's not illegal then it's legal - in British jurisprudence tradition at least https://en.wikipedia.org/wiki/Everything_which_is_not_forbid... - namely it's not law that impedes it (also in some places there's posthumous marriage)

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#104

Post by logicchains »

kybernetikos wrote: post #98
sillysaurusx wrote: post #38

Earlier quoted context omitted.

(I work on llama-dl.) We’re fighting back against the DMCA requests on the basis that NN weights aren’t copyrightable. This thread has details: https://news.ycombinator.com/item?id=35393782 I don't think you have to worry about Facebook going after you. The worst that will happen is that they issue a DMCA, in which case your project gets knocked offline. I don’t think they’ll be going the RIAA route of suing individu…

If NN weights aren't protected by IP law that could slow down progress quite a lot. That could be very good for people worried about alignment.

>If NN weights aren't protected by IP law that could slow down progress quite a lot.

What do you mean? IP law is overwhelmingly an impediment to progress; innovation happens faster when people are free to build on existing weights.

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#105

Post by muyuu »

gaogao wrote: post #33
0cf8612b2e1e wrote: post #12

Earlier quoted context omitted.

Not a lawyer, but that still feels like dubious territory. I would still be on the hook for acquiring the original download, which Facebook has been launching dmca takedown requests for the llama-dl project.

It's fairly similar to a ROM patch in the video game space, which has mostly stood the test of time.

like that but requiring 60GBs of CPU RAM for some reason :-P

one has to wonder how did they implement the storage of those deltas to require that sort of RAM

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#106

Post by yreg »

xbmcuser wrote: post #88

Just when you think Nvidia will go down something happens that changes it. These days unless you were into gaming or a machine learning dev the integrated graphics were good enough. But now first time in a long time I am interested in getting a gpu for running some of these chatbots locally.

As a very occasional gamer who uses an iMac for work I thought about getting a gaming PC for like 6 years.

Last fall it seemed that all the stars have aligned. The crypto winter and Ethereum switching to proof of stake meant that GPU prices fell to a reasonable level, I knew i would have a bit of a time to play some game during the holidays and as soon as Stable Diffusion was first posted on hacker news I knew that that's my excuse and my sign.

So far I think I have spent more time tinkering with the 20 python environments I have[0] for all the ML projects than playing RDR2.

[0] https://xkcd.com/1987/

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#107

Post by 7to2 »

gigel82 wrote: post #66

All of the llama derivatives are tainted by Meta's license, which makes commercial use illegal and even personal use dubious. Is all the training material used for Llama available as open source? Maybe lots of folks can pool their resources and create fully open clean models / weights instead.

> All of the llama derivatives are tainted by Meta's license, which makes commercial use illegal and even personal use dubious.

This is not true if you never agreed to Meta's license. If you haven't, you either can't redistribute the weights or you're completely free to use them as you see fit depending on whether weights are copyrightable (very likely) or not. We'll have to wait for the llama-dl lawsuit to find out for sure.

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#108

Post by muyuu »

ImprobableTruth wrote: post #74
sillysaurusx wrote: post #38

Earlier quoted context omitted.

(I work on llama-dl.) We’re fighting back against the DMCA requests on the basis that NN weights aren’t copyrightable. This thread has details: https://news.ycombinator.com/item?id=35393782 I don't think you have to worry about Facebook going after you. The worst that will happen is that they issue a DMCA, in which case your project gets knocked offline. I don’t think they’ll be going the RIAA route of suing individu…

I'm curious, do you not think this might have adverse effects? Namely, if NN weights aren't copyrightable, limited releases like Meta has done might not be possible anymore so they might just cease completely with releases, ultimately leading access to large models to be more restricted.

i honestly do not know what is worse from the three realistic alternatives:

1- to have large corporations and people with privileged access to them have these models exclusively and have them collaborate as a clique

2- to have those models openly released to everybody, or de-facto released to everybody as they leak in short order

3- to have the people who think releasing models is a bad thing simply not release them and work alone in their proprietary solutions, as the smaller companies and hobbyists do collaborate

i say let them have a go at number 3 and see how that works for them - shades of "Microsoft Network" vs Internet all over again

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#109

Post by stevenhuang »

gigel82 wrote: post #95

One of my favorite queries to assess an LLM is Is it legal for a man to marry his widow's sister? Vicuna-13B: The legality of a man marrying his widow's sister (widow marriage) varies depending on the country and cultural context. In some societies, widow marriage was traditionally practiced or tolerated. GPT-4: It is not possible for a man to marry his widow's sister, as the term "widow" refers to a deceased man's w…

Nice test, cool to see gpt4 got it.

You'd probably need to come up with a new one now though, or confirm knowledge cutoff for the next evaluation :p

Re: State-of-the-art open-source chatbot, Vicuna-13B, just released model weights

#110

Post by revelio »

chii wrote: post #80
anonymousgpt1 wrote: post #48

Earlier quoted context omitted.

Lemme save whoever is donating the legal here the time: model weights are definitely copyrightable.

> model weights are definitely copyrightable. on what legal theory or precedence makes this true? IMHO, the weights are akin to the list of telephone numbers in a directory - which is definitely not copyrightable; only the layouts and expressive portion of a phone directory is copyrightable. So to make the weights copyrightable, it needs to be argued that the 'layout' of the weight is a creative expression, rather th…

"Definitely" is too certain w.r.t. law, but it's pretty obvious how you'd argue these fall under copyright. The difficulty would really be the opposite, it'd be arguing the weights are not derived works of the copyrighted input data sets.

Firstly, weights are not merely a collection of facts like a telephone book is. If two companies train two LLMs they'll get different weights every time. The weights are fundamentally derived from the creative choices they make around hyperparameter selection, training data choices, algorithmic tweaks etc.

Secondly, weights can be considered software and software is copyrightable. You might consider it obvious that weights are not software, but to argue this you'd need an argument that also generalizes to other things that are commonly considered to be copyrightable like compiled binaries, application data files and so on. You'd also need to tackle the argument that weights have no value without the software that uses them (and thus are an extension of that software).

Finally, there's the practical argument. Weights should be copyrightable because they cost a lot of money to produce, society benefits from having large models exist, and this requires them to be treated as the private property of whoever creates them. This latter one should in theory more be a political matter, but copyright law is vague enough that it can come down to a social decision by judges.

Post reply on HN