Why use GPU with Neural Networks?

preview_player
Показать описание
Start with an analogy. Then delve into CUDA with some pytorch code to demonstrate why we use GPUs instead of just CPUs.

REFERENCES
Рекомендации по теме
Комментарии
Автор

Today I watched atleast 5 videos on differences between CPU and GPU, and this is by far the best one

addoul
Автор

It’s not just about memory bandwidth. It’s about the amount of arithmetic units in GPU, that can do tons of operations in parallel.

KlimovArtem
Автор

Analogies used in the explanations were really intuitive 👍

karthikragunathanandakumar
Автор

My man. May-trix. Mmmmaaaayyytrix. Matrix.

PS awesome vid

DeanCulver
Автор

Can you make a video about How to run jupyter notebook Using GPU in windows 11? Some suggested me to use Docker also, but I didn't find any video explaining how to use docker to run GPU powered jupyter notebook on windows 11 :(

sangeethajith
Автор

U should use trick like dual decoposition insted of GPU
because when u are using GPU windows will cost extra resources to allocate matrix to GPU
speed is not not guarantee in this step and Graphic card also very expensive including misc
comptaible botherboard CPU HDD
So I recommend dual decoposition

ccuuttww
Автор

why you don't make a series let's build a Robot
or series to make a hole program to learn a deep learning

ALG