neoneye2

9 ms·

31.
ARCathon 2023 winners announced. ARC = Abstraction and Reasoning Corpus(lab42.global)
1 points by neoneye2 3y ago|1 comments
32.
by neoneye2 3y ago
Both 1st place solutions solves 30 tasks of the 100 hidden tasks.
33.
by neoneye2 3y ago
Impressive how few code changes they managed to convert an existing LRU library into SIEVE. https://github.com/cacheMon/lru-rs/commit/1c45ef7fe85f4a4395...
34.
by neoneye2 3y ago
How can a dataset be formatted for fine tuning llama?
35.
by neoneye2 3y ago
Great video by CuriousMarc, about how the Apollo memory works. (Core Rope Memory) https://www.youtube.com/watch?v=hckwxq8rnr0
36.
by neoneye2 3y ago
It has 3 stars on github.
37.
by neoneye2 3y ago
ACE framework by David Shapiro. https://github.com/daveshap/ACE_Framework In the following video David explains the code and prompts. I built a thinking machine. Happy birthday, ACE! https://www.youtube.com&
38.
by neoneye2 3y ago
The front page of this website has a 3d demo'ish interactive ui. Love it. https://acko.net/ Great article.
39.
by neoneye2 3y ago
Tetris games with 4 cells, is that easier that Tetris games with 5 or more cells?
40.
by neoneye2 3y ago
If one were to make the ideal alternative to YouTube. What should the specs be like?
41.
by neoneye2 3y ago
Immediately after seeing werkkzeug, I started coding my own wz clone. Here is a video of what it was like around 2009. https://www.youtube.com/watch?v=ZPYy8lry2do
42.
by neoneye2 3y ago
Explanations of how werkkzeug works https://youtu.be/P7MNvUz5yyU?si=N1zED81UlYfuAV8N https://www.youtube.com/watch?v=uUfV41HE_Dg
43.
by neoneye2 3y ago
Has anyone run this on a Mac with M1 chip with GPU acceleration enabled? I did my first hello world fine-tuning of Llama2 today, using Google Colab and various pieces of code. So no resume from checkpoint, I had to specify the number of epo
44.
by neoneye2 3y ago
Cute videos about AI alignment here: https://www.youtube.com/@RationalAnimations IIRC some of them also involves politics and AI.
45.
by neoneye2 3y ago
AI "political" alignment is terrifying.
46.
by neoneye2 3y ago
Directory Opus, what makes it good? I have seen DOpus mentioned several times over the years. is it keybindings, features, performance? What is your impression of "Worker"? http://www.boomerangsworld.de/cms/wo
47.
by neoneye2 3y ago
Shameless self promotion of my own AI experiments. Around 110000 files checked out related to the OEIS database. The project is open source here: https://loda-lang.org/ I'm experimenting training a vision transformer w
48.
by neoneye2 3y ago
AI training data, can be thousands of files.
49.
by neoneye2 3y ago
Wow, just tried this out. Seems great. The filter files, just by typing in a few characters. Nice.
50.
by neoneye2 3y ago
I made a dual tab file manager 10 years ago for mac ( https://github.com/neoneye/newton-commander ). It's written in Objective C. Unlike other dual tab file managers, it use the spacebar for quicklook which I find m
51.
by neoneye2 3y ago
There are only a few example images. It could be interesting seeing more examples of good/bad predictions.
52.
by neoneye2 3y ago
Suggestion. Rename from `func` to `fn` so it's more aligned with Rust.
53.
by neoneye2 3y ago
Great video explanation The Hidden Geometry of Error-Free Communication - Another Roof https://www.youtube.com/watch?v=Tmx-v4FiP6I
54.
by neoneye2 3y ago
There is also https://www.windows93.net/
55.
by neoneye2 3y ago
From left to right is the `N` value. From bottom to top is the `a(n)` value. A342730 a(n) = floor((frac(e n) + 1) prime(n+1)). https://oeis.org/A342730 Knowing the N value and the previous nearby prime points, I guess
56.
by neoneye2 3y ago
A while ago I made a nice plot of how the prime numbers are distributed. Starting out somewhat wobbly, but getting more evenly spaced for bigger N's. Using irrational numbers and it would often yield amazing patterns. https:/&#x2
57.
Arc: Where do we stand today? (Abstraction and Reasoning Corpus)(lab42.global)
2 points by neoneye2 3y ago|0 comments
58.
by neoneye2 3y ago
ARC is "Abstraction and Reasoning Corpus". The ARC 1 dataset is here. https://github.com/fchollet/ARC The ARC 2 dataset is crowd sourced. If you can come up with a challenging task, then please contribute it.
59.
Last chance of contributing to the ARC 2 dataset, ends 30 June 2023(lab42.global)
1 points by neoneye2 3y ago|1 comments
60.
by neoneye2 4y ago
I have also made an interpreter using Rust. https://github.com/loda-lang/loda-rust/tree/develop/rust_pro... You can try the interpreter in the browser for computing the prime numbers: https:/&#x2F

More ›