Install Tensorflow GPU on Windows

I show all of the steps needed to install TensorFlow GPU on Windows. I setup an entire Machine Learning development environment as well by showing how to setup CUDA, cuDNN, Visual Studio Code, Jupyter and more.

LINKS MENTIONED

MY UDEMY COURSES ARE 87.5% OFF TIL December 19th ($9.99) ONE IS FREE

MORE DATA SCIENCE VIDEOS

#TensorFlow #Machine Learning

[yassine]

Yeah, tensorflow, finally.
I was waiting for this for so long.
I hope this is the first episode in a comprehensive series.
Thank you Derek my hero. 👍 respect

[jessmac]

TF doesn't support GPU anymore on versions >2.11. And I'm having a hard time trying how to install 2.10 version as it is not available thru pip method :(

[pitiwatlueang]

I baffled this for 1 DAY STRAIGHT!, thank you very much for providing this video. It very help me.

[BlackWolfy]

it doesnt work when I pip install tensorflow gpu :


Collecting tensorflow-gpu
Using cached tensorflow-gpu-2.12.0.tar.gz (2.6 kB)
Preparing metadata (setup.py) ... done
Collecting python_version>"3.7" (from tensorflow-gpu)
Using cached (3.4 kB)
Building wheels for collected packages: tensorflow-gpu
Building wheel for tensorflow-gpu (setup.py) ... error
error: subprocess-exited-with-error

× python setup.py bdist_wheel did not run successfully.
│ exit code: 1
╰─> [18 lines of output]
Traceback (most recent call last):
File "<string>", line 2, in <module>
File "<pip-setuptools-caller>", line 34, in <module>
File "/tmp/pip-install-ts44axh4/tensorflow-gpu_19edcea1be88430296c40c1d1f86831b/setup.py", line 37, in <module>
raise
Exception:


The "tensorflow-gpu" package has been removed!

Please install "tensorflow" instead.

Other than the name, the two packages have been identical
since TensorFlow 2.1, or roughly since Sep 2019. For more
information, see:



[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for tensorflow-gpu
Running setup.py clean for tensorflow-gpu
Failed to build tensorflow-gpu
ERROR: Could not build wheels for tensorflow-gpu, which is required to install pyproject.toml-based projects

[carloscastc]

Man, as well as the rest of the comments It worked! Thanks, really been struggling. Currently running my code while I write this comment and now that It runs on my GPU (Finallyyyy) it´s like 16 times faster. Best regards from Colombia!

[李墨-js]

You are excellent，I have been installing cuda for a completely night before seeing your video, after which I finally run successfully.👍

[Jerrel.A]

3:45, Is it possible to use Visual Studio Code instead of VS 2017 Community (because of CUDA v10.1-11.5)?

[antonio_cds]

thanks bro, works for me with tensorflow 2.9.1
, cuda 11.7 and cudnn 8.4

[derekcentineo]

Awesome video, best explanation on youtube mate.

[veczz]

bro thanks so much, I was about to pull my hair out

[xinchenzou]

You've really saved my life... Thank you so much for this helpful video!!

[bobjohnson]

waiting for tensorflow tutorial very excitedly

[AagneyWest]

Thankyou somuch man! I've been searching for a good tutorial on this for sometime. ❤️🤝 Grateful

[antonioaugello]

Guys if you get problems with tensorflow-gpu just use tensorflow==2.8.0, CUDnn 8.1 and CUDA 11.2

[MehdiSattari]

What was the purpose of installing Visual Studio? Where was it used? I think it is required if you want t compile and build TF from the source code

[BhaveshKumar-lkki]

No more cuda or cuDNN for tensorflows 2.11+

[unoclock]

hello sir，

Could i ask which version of CUDA&cuDNN is right for 1050ti?
it is an old Graphics card so i cant search the latest information.And i m not sure which is right.

THANKD A LOT!!!

[shivamraj]

Derek are there any before requirement before start learning this ML things...if there so tell me some base things for start with

[nicolaspujante]

I did all the steps with the correct versions and check the PATH but i still have this error "Could not load dynamic library 'cudart64_110.dll'; dlerror: cudart64_110.dll not found", I have a True return when i try to see if the built is with CUDA, but tensorflow can't see my GPU with that error. Anyone knows what can i do?, i have a rtx 3070, cuda 11.2, cudnn 8.1.0 and tensorflow-gpu 2.8

[karinemendes]

Thank you for the tutorial, it really helped!