Thank you!

This was great. I have seen countless videos on RNNs and LSTMs and nobody explained it by taking a simple example like you did. It was a perfect balance of theory with application.

That was incredible! thanks!

This is the best explanation of transformers models, please keep going on this channel. There are lots of models still need to explain!

For a neuroatypical, your background music is too distracting, it would be nice if there wasn’t any

thank you

You did a fine job.

Bro I use a setup with 128GB RAM with RTX 4080 16GB, RTX 3060 OC 12GB, RTX 2060 super 8GB on it with 5TB SSD M2 but I use as a open AIR setup for faster cooling. But you cabinet setup looks very efficient and cool. Could you send me the link for it

Woah! Exquisite, It's a 15 min video but I spent over an hour taking notes and understanding. You have done a great job, keep it up. Thank you so much! Such explanations are rare. ;)

Cool! 👏 Thank you for the video! What about the OS? Maybe Linux? If Linux, which one?

Una estructura árbol de la vida. Kabakah

12:56 isn’t that the other way around (i.e. they queries come from the previous layer and the keys and values from the encoder’s output)?

Excellent

I would like to build one but I'm way ignorant on coding and building.

it was not just help full but amazing, loved it.

Please create more videos. You are really good.

For a guy who did a "lot" of research and you want mores cores why not get a thread ripper.

thanks, bro. well explained.

How do i fine tune something?

😮

Hi, super explanatory and easy to follow video! Do you have any updates? Maybe using "cheap" AMDs like 6700xt? haha

i'm in love <3

The audio kept buggin me, I'd heard it somewhere, then I remembered the Iconic Outros of the Channel Veritasium !!!!!

in these figures, where are the weights?

🎯 Key Takeaways for quick navigation: 00:00 *Los transformers están revolucionando el procesamiento del lenguaje natural, superando a modelos anteriores como las redes neuronales recurrentes.* 02:29 *Los transformers introducen una arquitectura basada en atención, permitiendo un acceso potencialmente ilimitado al contexto durante la generación de texto.* 05:04 *La atención múltiple es un módulo clave en los transformers, que permite que cada palabra se relacione con otras en la secuencia de entrada.* 09:24 *La capa codificadora de los transformers utiliza la atención para crear una representación continua de la información de entrada.* 11:01 *Durante la decodificación, se aplica un enmascaramiento para prevenir que el modelo acceda a tokens futuros, asegurando una generación autoregresiva coherente.* 14:39 *Los transformers, al superar las limitaciones de la memoria a corto plazo, son especialmente eficaces para codificar y generar secuencias largas en el procesamiento del lenguaje natural.* Made with HARPA AI

Thank you so much! Very well Explained, cleared most of the doubts.

is it possible to build this on raspberry pi 3 plezzz reply 😇

Can we use CNN instead of RNN here for the classification of MFCC images?

bro you made this video before gpt 3 and all the new era of LLMs !!

best explanation with good visualization

This is gold, man. Thanks!

Excellent , I got a clear explanation

The i in positional encodings is an index to the dimension of the embedding, not the "timestep"

Chad Explanation

not much

thanks a lot, it was a lot of info but very helpful

where can i find the code of the system ?

看進步健康情形不僅是看不見卡巴斯基開心吧就是

I have a spare 3080 and want to build a deep learning machine/home server PC. Probably wont be that fast, but I figure it will be good enough for me to further my education.

Did u use GPU for training this

Hi, many thanks for the great video. One quick question, if I choose AMD Ryzen, would I be able to install CUDA suit on it?

Why does the decoder select the token with the maximum probability instead of randomly selecting a token based on the probability distribution?

You have such a sweet and pleasant voice. Thank you, mate, for the good explanation. 😊

super helpful. i want to be an ai engineer as well.

This explanation is INCREDIBLE!!!

great explanation

bro gave apple the pass lol worst of all

we allready done this decade ago. and it depend on language. if you write and spell same way then its easy but if...english then LOL. different woords spelled same! lol