emadm

8 ms·

61.
by emadm 3y ago
Aleph Alpha raised even more ^_^ https://sifted.eu/articles/ai-startup-aleph-alpha-raises-500...
62.
by emadm 3y ago
They already do, we just released a model equivalent to most 40-60b base models that runs on a MacBook Air no problem. It's like 1.6gb, ones coming are better and smaller https://x.com/EMostaque/status/1732912
63.
by emadm 3y ago
Yeah they shouldn't worry, they'll get a big French government deal at worst
64.
by emadm 3y ago
As an open multi modal we went for a flat membership scheme we are introducing: https://x.com/EMostaque/status/1729609312601887109?s=20 $1 => $100k+ flat for all models we release depending on company size&#x2F
65.
by emadm 3y ago
Mistral doesn't release source code or data just weights. They did include some details of sliding window attention previously
66.
by emadm 3y ago
Know it well from video game days, but its being set up for huge predictability / stability given we have simple company control. The base membership is even separate from other services, designed to support open model development.
67.
by emadm 3y ago
Nah its flat fee for all the models, no revenue share. Amazon Prime but for generative AI, large companies have signed up for huge deals already and everyone has been happy with pricing, its designed to be simple base then we upsell other s
68.
by emadm 3y ago
Yeah got a way to beat 3.5 but it beats most of the first generation llama tunes even guacano 65b Lots of improvements to go
69.
by emadm 3y ago
We included full training details for the base model on 4 trillion tokens including wandb etc https://stability.wandb.io/stability-llm/stable-lm/reports/S...
70.
by emadm 3y ago
We looked at that, will be self service for commercial with flat pricing including all base models, weights are all downloadable by anyone. Models are very interesting
71.
by emadm 3y ago
It’ll be static with cpi max, flat membership for all core models Just released video, sdxl turbo and 3d, code and more coming Very positive reaction so far and we will still do our grants for OSS and do OSS collaborations, done over 10m A1
72.
by emadm 3y ago
It’s about twice the speed
73.
by emadm 3y ago
It will be included under our membership next week which starts at $1 a month after grant ($20 base)
74.
by emadm 3y ago
We build the best video, image and other models with more downloads and usage than anyone. It is quite revolutionary for creative industry which is a few hundred billion in size, which is a reasonable market globally.
75.
by emadm 3y ago
Check out our fully open recent 3b model which outperforms most 7b models and runs on an iPhone/cpu, fully open including data and details Tuned versions outperform 13b vicuña, wizard etc https://stability.wandb.io/stab
76.
by emadm 3y ago
We back rwkv, eleuther ai and others at stability ai We also have our carper.ai lab for the rl buts We are rolling out open language models and datasets soon for a number of languages too, see our recent Japanese language models for example
77.
Stable LM 3B full technical report(stability.wandb.io)
4 points by emadm 3y ago|0 comments
78.
by emadm 3y ago
It's a model 40% of the size of Mistral's designed to be transparent (full training details, datasets & evals here: https://stability.wandb.io/stability-llm/stable-lm/reports/S... ) and work on e
79.
Stable LM 3b4T, 20B LLM performance in 3B(stability.wandb.io)
4 points by emadm 3y ago|2 comments
80.
by emadm 3y ago
Suppose smol language model
81.
by emadm 3y ago
It’s coming
82.
by emadm 3y ago
Carper is technically a Stability AI lab and community https://twitter.com/carperai Core are full time Stability AI
83.
by emadm 3y ago
We fund Eleuther AI too as it says there (but made sure it can run independent when 501(c)3 set up) and provided compute for RWKV. Conjecture was meant to be alignment so can't really support, stability scaled resource provision.
84.
by emadm 3y ago
It has good potential and nobody ever thought that RNN's would be able to scale as they have led by BlinkDL. Well we did at Stability AI which is why we provided and scaled compute for it. They have the potential for far better edge in
85.
by emadm 3y ago
The CUDA moat argument just doesn’t hold up in real life for foundation model training and inference We get equivalent performance on non-NVIDIA chips without it and most of the stuff is abstracted away these days We do write CUDA when need
86.
by emadm 3y ago
He is indeed the longest serving CEO in Silicon Valley More impressive to me is that he didn’t let folk go because of the many downturns on the way
87.
by emadm 3y ago
“ I have a hard time seeing what companies are going to go through the time and effort to port things that already run on CPUs to GPUs;” I mean someone will build an AI to do that right ^_^ Video and 3D are pretty much in the next year, all
88.
by emadm 3y ago
Not really I always used MA (Oxon) as appropriate in my resumes and similar. You don't call it a BA. I think the thing here is also intention given this is a weird thing, it would make 0 difference in anything I do for me to have tried
89.
by emadm 3y ago
It's pretty clear cut as his suit has three claims 1. Did not inform him of pivot and generative AI art (false, even generated AI art for him) 2. Did not inform him of fundraise/inflows (also false) 3. Either of the above are not
90.
by emadm 3y ago
That was just the hate mob attacking I noted that the most interesting data is behind firewalls as in private data as our entire model is open models to private data => not using that data to train our models You can also license that da

More ›