Thank you, very informative and clear, one of the most intuitive ones seen on YT, keep up the great job!
@grandson_f_phixis9480
3 күн бұрын
Thank you very much
@tsclly2377
7 күн бұрын
Did you burn up your SSDs? I did when ETH mining.. SLC is the only way to go and that means Optane (pedabytewrites), as the PCIe linked NVMe for the data dump (write) as it is as fast as the PCIe 4 bandwith (4x) and the lanes are really only 8x as 16x is 2 8x lanes
@zix2421
7 күн бұрын
Thank you, LSTM is really useful thing, I’ll use it
@archlunarwolf
7 күн бұрын
He sticks it to Amazon and then goes ahead and orders the SD card from Amazon...
@RoonyKingXL
7 күн бұрын
6:30 - Am I being stupid or is the visualization wrong? The scores matrix should be 3x3 not 4x4, right? Please correct me if I'm wrong, I feel like I'm missing something.
@Mojo522
13 күн бұрын
Thank you!
@prabhdeepsingh8726
13 күн бұрын
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.
@hamidrezahosseinkhani5980
15 күн бұрын
That was incredible! thanks!
@cocoph
21 күн бұрын
This is the best explanation of transformers models, please keep going on this channel. There are lots of models still need to explain!
@BrianCarter
24 күн бұрын
For a neuroatypical, your background music is too distracting, it would be nice if there wasn’t any
@Boom-em1os
24 күн бұрын
thank you
@markthomas2436
26 күн бұрын
You did a fine job.
@phoenix1799
28 күн бұрын
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
@tabindahayat3492
29 күн бұрын
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. ;)
@alexdaniel76
Ай бұрын
Cool! 👏 Thank you for the video! What about the OS? Maybe Linux? If Linux, which one?
@anamariatiradogonzalez
Ай бұрын
Una estructura árbol de la vida. Kabakah
@anthonybernstein1626
Ай бұрын
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)?
@guilhermealvessilveira8938
Ай бұрын
Excellent
@timothyweakly2496
Ай бұрын
I would like to build one but I'm way ignorant on coding and building.
@ashishbhong5901
Ай бұрын
it was not just help full but amazing, loved it.
@AravindUkrd
Ай бұрын
Please create more videos. You are really good.
@Sabumnim666
Ай бұрын
For a guy who did a "lot" of research and you want mores cores why not get a thread ripper.
@coolStranger516
Ай бұрын
thanks, bro. well explained.
@sweatyninja9755
Ай бұрын
How do i fine tune something?
@ayanah4821
Ай бұрын
😮
@martinsenuy895
Ай бұрын
Hi, super explanatory and easy to follow video! Do you have any updates? Maybe using "cheap" AMDs like 6700xt? haha
@walloouu
Ай бұрын
i'm in love <3
@aneekeshkumar8199
Ай бұрын
The audio kept buggin me, I'd heard it somewhere, then I remembered the Iconic Outros of the Channel Veritasium !!!!!
@sahhaf1234
Ай бұрын
in these figures, where are the weights?
@josep1429
Ай бұрын
🎯 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
@yashgajjar4838
Ай бұрын
Thank you so much! Very well Explained, cleared most of the doubts.
@chriz__3656
Ай бұрын
is it possible to build this on raspberry pi 3 plezzz reply 😇
@hussainbhavnagarwala2596
Ай бұрын
Can we use CNN instead of RNN here for the classification of MFCC images?
@user-wm8hy8ce2o
2 ай бұрын
bro you made this video before gpt 3 and all the new era of LLMs !!
@azharkhan-tr1wj
2 ай бұрын
best explanation with good visualization
@mohamedibrahimbehery3235
2 ай бұрын
This is gold, man. Thanks!
@sj5558
2 ай бұрын
Excellent , I got a clear explanation
@seangai126
2 ай бұрын
The i in positional encodings is an index to the dimension of the embedding, not the "timestep"
@muzammilomarzoy6616
2 ай бұрын
Chad Explanation
@charlesstevens3297
2 ай бұрын
not much
@abdulazizyaser1616
2 ай бұрын
thanks a lot, it was a lot of info but very helpful
@adambenghoula
2 ай бұрын
where can i find the code of the system ?
@tsunningwah3471
2 ай бұрын
看進步健康情形不僅是看不見卡巴斯基開心吧就是
@jasonjennings8465
3 ай бұрын
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.
@amruth2545
3 ай бұрын
Did u use GPU for training this
@wasifmasood969
3 ай бұрын
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?
@csvegso
3 ай бұрын
Why does the decoder select the token with the maximum probability instead of randomly selecting a token based on the probability distribution?
Пікірлер