Tutorial 33- Installing Cuda Toolkit And cuDNN For Deep Learning

preview_player
Показать описание
Please join as a member in my channel to get additional benefits like materials in Data Science, live streaming for Members and many more

Please do subscribe my other channel too

Connect with me here:

Рекомендации по теме
Комментарии
Автор

Mate! Thank you so much for this video! It was really helpful, I was following a tutorial from an online course but it didn't work and I was stuck. Your video showed me what the issues were and helped me fix them. All the best from Australia!

chrisber
Автор

If we want to use GPU only through Anaconda then this task of SETUP can be done in a single line of code ->
conda install -c anaconda tensorflow-gpu
It will automatically install all the dependencies like CUDA & CUDNN itself.

AmanSharma-ljff
Автор

CUDA is not just deep learning, what I love about, parrell universes always rocks.

satyam
Автор

Hi Krish Naik,

You are doing an impeccable job. I tried a lot of youtube videos which are slightly different compared to your but none worked. To be specific there are differences in handling cuDNN files.

Thank you for the video.

However, I found one thing out in the setup process.
We don't need to install C++ mobile and other development packages while setting this up. The basic installation of VS 2017/2019 can do the job. I went for 2019. I am writing this because this can save a lot of time and data charges for the user because it's a colossal amount of data i.e., 14Gigs almost 5 times more than all the other required files aggregated.

Thank You.

:) HAPPY CODING :)

bhanupalagati
Автор

(13:55 )Sir by mistake you have said wrong i think, when we are giving import tensorflow command, it will show error if in our environment we have not included bin of cuda which we dnt need to paste in C drive I think.
In environment we need to go to path as u guided and have to add " bin " present inside "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin" . This worked for me. I may be wrong. Pardon me If wrong.
I checked by going back in your video, "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin", already in path in your case.

while import tensorflow, it needs cudart64_100.dll and in cuda/bin we have cudart64_7.dll, so was giving error in my case. Hope this will help sm1.

rajatjaiswal
Автор

Hi Krish, I am new to Deep learning and your videos are really helping me alot. thank you so much. Keep making such great videos

suvarnamaji
Автор

Instead of adding System variable , You can also replace Files From cuDNN in the Direcory where Cuda is installed.
You will find Directories with same name there. Just Copy paste files to respective direcory.

ganeshkharad
Автор

Q: Other than cost, Any reason why you did not go with the Nvidia RTX GPU like the 2060? .. which has tensor cores which are proven to reduce deeplearning training significantly..

madhankumar
Автор

I was struggling with enabling GPU with GTX 1650. This session finally helped me installing CUDA correctly and happy to see GPU running the epochs fast ! Thanks very much :)

Raaj_ML
Автор

Great video thanks a lot. Others do not show the full process from scratch. Great content

noufbouf
Автор

Hey Krish, please let me know the correct procedure to install tensorflow in windows 10 PC with Python 3.7 installed or please guide me to a video if already published by you. There are so many confusing stuff about it out there and I am looking for the correct procedure. Thank you.

aniacharya
Автор

Thank you so much. I forgot and got stuck on the Cudnn part >< This saved me hours

ChadALozano
Автор

@13:17
Python 3.9.16 (main, Mar 8 2023, 10:39:24) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow
>>> import tensorflow
>>>import tensorflow

I m getting this only.

ashu
Автор

I installed visual studio 2017 without workloads and it works just fine

nehapant
Автор

really appreciate! finished installing

xingnanzhou
Автор

If you want to install tensorflow-gpu
Install anaconda, then create environment and activate, then
Conda install tensorflow-gpu
And its done

rohitarora
Автор

will it work with cuda version 10.1 and VS 2019?
in the tensorflow site it is mentioned that the TF is compatible with cuda 10.1

nasiksami
Автор

Sir, after installing tensorflow i am getting this error.
2020-04-21 07:05:44.235184: W Could not load dynamic library 'cudart64_100.dll'; dlerror: cudart64_100.dll not found

gauravsingh
Автор

Sir can you upload the video, how to do with pytorch.

khanwaqar
Автор

This was Great Help!!

The 1650 card is not listed in documentations and I was scared that I might have to return my laptop. I was relived when you started with "New Laptop with 1650 card". If you can add in the description that this is for 1650 gpu then it'll be of great help for people like me!

Anyways thanks for the help!!

amitRuns