GitHub – dabochen/spreadsheet-is-all-you-need: A nanoGPT pipeline packed in a spreadsheet
Posted by jpluimers on 2024/06/12
A great visualisation that LLM are basically a bunch of numbers: [Wayback/Archive] GitHub – dabochen/spreadsheet-is-all-you-need: A nanoGPT pipeline packed in a spreadsheet.
It also shows you that Excel is an excellent tool for working with numbers and formulas on a larger scale.
(note the file is a .numbers file developed in the Mac version of Excel)
Via:
- [Wayback/Archive] /Fay-lee-nuh/ on X: “Programmers: Spreadsheets aren’t code @chendabo: Hold my beer”
- [Wayback/Archive] Dabo on X: “I recreated an entire GPT architecture in a spreadsheet. It is a nanoGPT designed by @karpathy with about 85000 parameters, small enough to be packed into a spreadsheet file. It is great for learning about how transformer works as it shows all the data and parameters going”
I recreated an entire GPT architecture in a spreadsheet.
It is a nanoGPT designed by @karpathy with about 85000 parameters, small enough to be packed into a spreadsheet file.It is great for learning about how transformer works as it shows all the data and parameters going through a transformer pipeline, and all the calculations are actually working inside those cells.
Here is a link to the project where you can download the file and play with it yourself. No coding needed, it is just a spreadsheet. 😉
Spreadsheet is all you need:
github.com/dabochen/spreadsheet-is-all-you-needAlso thanks to @BrendanBycroft with his LLM visualisation project which inspired this.
[Wayback/Archive] cdn.syndication.twimg.com/tweet-result?id=1799614807878095359&token=!
[Wayback/Archive] GPmDaSBbsAAbhXl.jpg (1200×692)
[Wayback/Archive] GPmDaSAbMAEhbz2.jpg (451×1199)
--
jeroen
Sorry, the comment form is closed at this time.