Lesson 1: Demystifying how GPT works: From Architecture to...Excel!?! 🚀

Подписаться 2,1 тыс.

Просмотров 53 тыс.

50% 1

Think AI is too complex? Think again! If you're spreadsheet-savvy, you're ready to grasp modern AI.
Welcome to the first episode of a deep dive into how the brains 🧠✨ of modern AI (Large Language Models and Transformers) actually work.
In this 10-minute video we kick things off by:
🏗️ Walking through the high-level architecture of GPT-2.
📊 Witnessing the Transformer's phases come to life in an Excel spreadsheet.
👀 Stay tuned for our upcoming episodes where we’ll dive deeper, dissecting each phase of the Transformer step by step.
🔗 Helpful Resources:
Learn more and download the Excel sheet at spreadsheets-a...

Опубликовано:

21 сен 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 51

@RamiGB 6 месяцев назад

I am sorry when It sounds like a cliche but honestly this video deserve way more views! so people can understand how GPTs work! excellent beautiful work thank you.

@Spreadsheetsareallyouneed 6 месяцев назад

Wow, thank you. Please spread the word. A better understanding of AI is helpful as we have these societal debates about the technology.

@jinyoungkim9087 6 месяцев назад

Normally, I dont write comment but, this is the most useful video!

@Spreadsheetsareallyouneed 6 месяцев назад

Thank you for taking the time to write a positive comment!

@carlt.8266 6 месяцев назад

Thank you for explaining "Temperature" so easily

@jytou 6 месяцев назад

Excellent! Spreadsheets deserve more attention! 😀

@rajahx 6 месяцев назад

Bravo! This is great way to introduce the nuts and bolts of large language models to a lay audience. Far, far more accessible than the baggage that comes with teaching it via pytorch or tensorflow. Just simple math and a little bit of excel wizadry. Lovely idea and thank you for making these videos.

@Seikou27 6 месяцев назад

This input is so DN awesome. Well done! Definitely deserves more attention :-)

@Spreadsheetsareallyouneed 6 месяцев назад

Much appreciated!

@lewisbentley3177 6 месяцев назад

This is fantastic! I'm really looking forward to the rest of the series.

@Spreadsheetsareallyouneed 6 месяцев назад

Thanks. Notes like this are helpful motivation to keep going.

@cz8986 6 месяцев назад

Best explanation I'have ever seen on transformers. Big KUDOS !

@Spreadsheetsareallyouneed 6 месяцев назад

Thanks!

@Vesture23 5 месяцев назад

Very instructive video, easily understood. Look forward to you future videos!

@paulwillis88 6 месяцев назад

Fantastic work! Thank you for the comprehensive and useful explanations

@Spreadsheetsareallyouneed 6 месяцев назад

Glad you enjoyed it!

@Cal1fax 6 месяцев назад

Really well done! Great explanations. I really hope you finish the series

@Spreadsheetsareallyouneed 6 месяцев назад

Me too...Putting these together has become more work than I expected but the positive reaction has been motivating.

@richardgrosman5798 6 месяцев назад

Thank you for simplifying!

@Spreadsheetsareallyouneed 6 месяцев назад

You're welcome!

@tarrychang7031 6 месяцев назад

hey, this project is great. Super interesting and educative. I shared with a lot of my friends. Thank you for your contribution!

@Spreadsheetsareallyouneed 6 месяцев назад

Thank you for sharing!

@nevermnd001 6 месяцев назад

Haha.. This is great/intructive, but also crazy... In Excel (!?!)-- I like it.

@Spreadsheetsareallyouneed 6 месяцев назад

Someone on Twitter put it similarly: "it is simultaneously incredible, and an aberration" twitter.com/bdkoepke/status/1769128618486112448

@j83lin 6 месяцев назад

Really educational video, thank you!

@Spreadsheetsareallyouneed 6 месяцев назад

Glad it was helpful!

@carlt.8266 6 месяцев назад

It is an amazing lecture you compiled in this video. Do you also have one on the training of GPT2 or how the tokens get their semantic context/meaning?

@GaryKeramidas 6 месяцев назад

Great job. noticed on the tokens to text embedding sheet you have embedding missing the "b" in row 14

@Spreadsheetsareallyouneed 6 месяцев назад

you're right. good catch! i'll have to fix that in subsequent versions.

@frankenmizer828 6 месяцев назад

Well done! Thank you!

@Spreadsheetsareallyouneed 6 месяцев назад

Thanks!

@tagkaz 6 месяцев назад

easy to understand. great!

@mrhaoji 6 месяцев назад

amazing! easy to understand for the beginners. BTW, can u share the Powerpoint too?

@change2023now 6 месяцев назад

It's in the description.

@mrhaoji 6 месяцев назад

didn't find the slides in the video besides the Excel file@@change2023now

@Spreadsheetsareallyouneed 6 месяцев назад

At some point I'm planning to turn them into blog posts which will likely include images from the powerpoint. If I did that, would it help your use case?

@mrhaoji 6 месяцев назад

sure@@Spreadsheetsareallyouneed

@kynamnguy 6 месяцев назад

So good brooo

@Spreadsheetsareallyouneed 6 месяцев назад

Thanks!

@GrossGeneralization 6 месяцев назад

🤯

@fastica 6 месяцев назад

What I've learned from this video: Mike is quick and he moves quickly.

@Spreadsheetsareallyouneed 6 месяцев назад

Lol. You have a bright future as a Large Language Model. 😃

@ricp 6 месяцев назад

this is amazing!

@fxpatrickryanto2452 6 месяцев назад

thank you for great explanation and the spreadsheet file, yet talking about v0.6.0, would you mind to explain where is the Split_Into_Characters formula is located? because when i run the manual calculation, sheet Prompt_to_Tokens, cell (column 2, row 4) only return #NAME? error

@TNIxeye 6 месяцев назад

same here

@robinhodson9890 6 месяцев назад

I haven't downloaded the sample spreadsheet, but he notes in a different video on tokenisation, that it splits words into one or more tokens, not one token per word, and certainly not one per character, as these would result in a drastically larger memory footprint.

@Spreadsheetsareallyouneed 6 месяцев назад

This could be locale issue or an Excel version issue. See github.com/ianand/spreadsheets-are-all-you-need/issues/6

@-iIIiiiiiIiiiiIIIiiIi- 6 месяцев назад

BIG MIKE!

@SeaJay_Oceans 6 месяцев назад

0:23 so... 2030 Sapiens Silicon ?

@Spreadsheetsareallyouneed 6 месяцев назад

I've been waiting for someone to comment on that...It was a funny throwaway observation right after I put the slide together but who knows. Ray Kurzweil has been predicting AGI by 2029 for over two decades now.

@SeaJay_Oceans 6 месяцев назад

@@Spreadsheetsareallyouneed We Can Think, Therefore, We Evolve ...