today at 9:32 PM
Isn't part of the Jev marketing that it has "terra-class intelligence"? I don't know how much it actually achieves that, but unless that's EXTREMELY wrong, it's hard to see how a 0.3B model could claim to be an OS Jev.
today at 9:44 PM
agree, maybe os jev is too strong of a wording.
when more people get the idea how to create system one, I am sure Laya will have lots of use cases.
today at 9:39 PM
So cool. I've fired up pumas (energy monitor) and it seems to run almost fully on the neural engine and not the GPU so it plays really nicely with CoreML
today at 7:59 PM
Based on my admittedly limited research, it seems like you should use Laya for much more deterministic tasks where you have some training data. It won't be as good as Jev for zero shot cases.
today at 8:11 PM
I think it can make a lot of sense to make a set with jev and then train laya on that set and do the rest with laya for saving $
today at 8:08 PM
agree, too good to be true for one shot cases
today at 5:12 PM
Local LLMs are the future, and one of the reasons I think the data centre furore is just going to end in a market crash.
LLMs that can reliably be used for control problems are the future, and I think classic/deep RL has generally been overlooked for years for a whole host of problems by wider industry because it felt inaccessible. The first thing I thought of when I saw Jev (and then Laya), was "this might move the needle in a really, really interesting way".
Local LLMs that can reliably be used for control problems smash through a lot of barriers I'm interested in, and this intrigues me a lot. Guess I'm about to become a big Laya fan if it can run on this kind of hardware to this performance.
today at 7:20 PM
The future for whom? The general public? Not a chance, no way, not unless it's able to run on a phone (anywhere from 20-40% of internet users, world-wide, are phone-only).
For companies? I think that's a lot more plausible, as that's mostly just a question of money - is it cheaper to run and administrate our own models, or outsource that?
For technically inclined users? I think that's unlikely unless they're able to operate on relatively cheap hardware while still being just as good as the hosted models. And by that I don't mean "a mac studio," that's far more money than I think is reasonable. A single RTX 5080, maybe, once memory prices start to drop.
today at 7:57 PM
Compare the games your average high-end smartphone can run to the AAA titles of the 2010's. It's not a matter of "unless it is able to" but "when it is able to".
today at 5:20 PM
that's not a local LLM. If it's local, it doesn't matter in this case. Laya is a System 1 "AI", namely works like a classifier, given a state and questions, it shoots probabilities for each. I publish an episode tomorrow about Laya and Typesafe AI on https://www.youtube.com/@DataScienceatHome
Stay tuned ;)
today at 7:41 PM
cool, will check it
today at 5:44 PM
It won't. Laya is a finetuned version of Google's BeRT model, which is almost 10 years old right now.
If BeRT had any potential to disrupt the datacenter buildout, it already would have.
today at 9:08 PM
Modernbert is from 2024. It's also trained from scratch, not a fine tune.
today at 7:12 PM
How much memory does this use of the test machine's (M3 Max) 128 GB unified memory?
today at 7:34 PM
The model is only 0.3B params so probably not much.
today at 8:07 PM
Physical footprint: 560.4M Physical footprint (peak): 778.0M
today at 5:21 PM
Great job! Did you finetune your own Laya for the snake game or what?
today at 8:27 PM
its just an example how to run it locally. it takes the laya from official repo with the provided snake example. you can try running it in 2 minutes
today at 5:17 PM
This looks like a local AI model playing Snake -- is that correct? The article offers no explanation.
today at 7:37 PM
Correct, the cli commands are just copy paste to run on your machine.
today at 5:19 PM
The linked github project [1] contains more information.
[1]: https://github.com/mizorewww/laya-coreml
today at 5:15 PM
[dead]
today at 6:18 PM
[dead]