d3m0t3p

7 ms·

1.
by d3m0t3p 11d ago
They are currently targeting more material science, such as quantum chemistry for drugs discovery or new material discovery. They can’t and do not aim to do prime factorization on those.
2.
by d3m0t3p 11d ago
I agree that they should be doing a press release and a separate technical deep dive. Here they mix both and don’t get the public. It has always been IBM problem to correctly target the product to the audience. We see it here again
3.
by d3m0t3p 2mo ago
When everyone push on friday, and you have 400 CICD pipeline triggers spawning that many nodes. How do you know if this is unexpected ? Their cloud compute might be on demande, someone starts training a model and 50 machines are spawned. Kn
4.
by d3m0t3p 3mo ago
They are using Qwen, so this is decoder only.
5.
by d3m0t3p 3mo ago
This is exactly the same with providers from the USA.
6.
by d3m0t3p 3mo ago
It was 15 years ago. Whole management got replaced, they are quite ambitious. Let's see if how this works out now.
7.
by d3m0t3p 3mo ago
I don't think so, faster trains are overtaking slower trains. There is simply not enough space between the station to overtake without having an acceleration that would damage the trains or the tracks. For example in western switzerlan
8.
by d3m0t3p 4mo ago
Interesting to see clearml but not its bigger counterpart mlflow
9.
by d3m0t3p 4mo ago
I think this is due to their AI insight, they run locally a model and it start to burn the whole computer.
10.
by d3m0t3p 4mo ago
It is really fun that the navbar has unaligned elements. (Docs is lower)
11.
by d3m0t3p 8mo ago
Yea but the goal it not to bloat the context space. Here you "waste" context by providing non usefull information. What they did instead is put an index of the documentation into the context, then the LLM can fetch the documentati
12.
by d3m0t3p 9mo ago
Same, Firefox iOS
13.
by d3m0t3p 9mo ago
The model is fined tuned for chat behavior. So the style might be due to - Fine tuning - More Stylised text in the corpus, english evolved a lot in the last century.
14.
by d3m0t3p 9mo ago
Is that really the only thing you managed to remember ?
15.
by d3m0t3p 11mo ago
Because the ML ecosystem is more mature on the NVidia side. Software-wise the cuda platform is more advanced. It will be hard for AMD to catch up. It is good to see competition tho.
16.
by d3m0t3p 11mo ago
In my own studies, software engineering was mostly about structurig code, coding pattern such as visitor, singleton etc. I.E how to create a maintainable codebase
17.
by d3m0t3p 1y ago
Would you have some literature about that ?
18.
by d3m0t3p 1y ago
This sounds a lot like what the Muon / Shampoo optimizer do.
19.
by d3m0t3p 1y ago
Interesting to see that they enforce retroactive opt out for data collection. I wonder how they do that, what if the model is already trained with your data and you opt out.
20.
by d3m0t3p 1y ago
You can batch only if you have distinct chat in parallel,
21.
by d3m0t3p 1y ago
Nice to see a master thesis highlighted on the research groupe page
22.
by d3m0t3p 1y ago
Your first link is (in my opinion) highly biased in the samples they choose, they hired maintainers from open-source repos (people with multi years of experience, on their specific repo). So indeed, IF you are in that case: Many years on th
23.
by d3m0t3p 1y ago
It is Gemma 3n, I can't give feedback yet on the battery hit, But I would not expect anything bad as these models have been developed for much smaller devices (Phones)
24.
by d3m0t3p 1y ago
Hey, really cool project, I’m excited to see the outcome. Is there a blog / paper summarizing how you are doing it ? Also which research group is currently working on it at eth ?
25.
by d3m0t3p 1y ago
Do they ? LLM embedd the token sequence N^{L} to R^{LxD}, we have some attention and the output is also R^{LxD}, then we apply a projection to the vocabulary and we get R^{LxV} we get therefore for each token a likelihood over the voc. In t
26.
by d3m0t3p 1y ago
You can check company names too ! It's interesting to see that by default, the graph shows google,apple. But adding meta, and IBM really changes the plot. Meta went from 2K to 10K+ from 2018 to 2025. While IBM seems to have stopped con
27.
by d3m0t3p 1y ago
Apparently this is the case: https://github.com/tulir/whatsmeow/discussions/199
28.
by d3m0t3p 2y ago
Hi, do you offer visa / allow remote from the EU (GMT+2)
29.
by d3m0t3p 2y ago
Would you mind sharing your company name? I'm a master's student in AI, and after finishing my master's thesis at IBM this summer, I'll be looking for jobs.
30.
by d3m0t3p 2y ago
Well, it's just like stochastic gradient descent, if you think about it. The normal gradient descent is computed using the whole training set. The stochastic gradient is trained on a batch (a subset of the training set), and in the dis

More ›