fzimmermann89

The switch by Artificial Analysis from per-token-cost to per-benchmark-cost shows some effect! Its nice that labs are now trying to optimize what I actually have to pay to get an answer - It always annoys me to have to pay for all the senseless rambling of the less-capable reasoning models.

..and for complex valued tensors, you need to conjugate.

Also, for an auto complete I think a small llm trained from scratch should already work well. Have you tried on if the tinystories(also only 3gb..)/nanogpt speed runs without any fancy loss terms etc as a baseline?

How foreign is the language - was it likely included in pre training to some degree? Does it use grammar, syllables, and logic similiar to one of the large languages? Your approach assumes there is an easy to learn mapping between context in your target language and concepts in a prettained llm.

Can you get more text written in the low resources language?

Are you ok to share the name of the language?

Contacting support obviously interfered with Apple services. Duh.

If I am not mistaken, this is done by modulation in Fourier space. We have already been using this in optical setups for ages - at the speed of light.

The interesting part imo is the implementation of this idea in their work and the efficiency and physical size.

Sad to hear that they removed most of their content as well.

If only they would fix the memory leak and freeze on resume from hibernate that has been an issue for the last year at least...

I thought the same, until I noticed a really annoying WSL2 bug: On two machines I own, waking up from hibernate or standby causes a wsl related process (vmmem) to consume 100% CPU, with wsl becoming completely unresponsive (including wsl terminate etc).

You have to kill all wsl processes, which requires admin rights. So without elevated rights, Ubuntu on windows is not usable on these laptops.

The issue is known for years and has hundreds of comments on GitHub without a fix (https://github.com/microsoft/WSL/issues/6982)

*shoes.

I got access by providing an academic email adress without mentioning any relevant publications etc.. Took maybe 2-3 days..

There is a typo in the in app purchase page: "unlimmited"

I clicked on one story, did not like it, so i canceled it. I wanted to give the app a second chance. But alas, only one story is allowed per day.. Maybe increase this at least for the first day or do not count unfinished stories....

So the code looks (apart from pointers) similiar to numba which feels much closer to numpy/pytorch high level code. Are there huge advantages in the triton model compared to numba that I don't see? Or is there a big performance gap? For me numba was always the easiest way to get some new idea running on cuda, and most of the time it was fast enough..

Did anybody find performance comparison between numba and triton?

or the text went through latex at some point, which can easily lead to a"->ä.

letter+quotation mark is a common way to write umlauts...