Install TensorFlow-GPU 1.8 on Windows PC

Hassle-free step-by-step guide to install tensorflow-gpu version 1.8 on computer with Windows operating system. Includes steps for installation of CUDA toolkit and cuDNN as essential pre-requisites for tensorflow.
Useful Links:

Please comment below for suggestion and/or queries.

[theos-]

Thank you a lot it works just fine!! Guys if you have any problems check which files you have downloaded. Make sure that you have downloaded the correct ones!

[ajitadash]

Thanks a lot!!! Finally after trying since 2-3 weeks, now I have successfully installed it.All thanks to you.Keep on doing this work...Kudos!!!

[rajatjain]

After all many attempts, I was able to install it. Thanks dude.

[francescopiemontese]

Only tutorial that worked for me, thank you so much

[KeterThree]

awesome! thanks! apparently the installation procedure has been considerably simplified for this version.

[DeveloperTharun]

THanks a Lot! It always takes me a day to install tensorflow-gpu though,
If it is still not working for anyone, then try uninstalling & deleting all the files related to tensorflow & cuda,
Then restart the PC, and try following this video, it will work for sure.
Cool

[NB]

nice video the instructions worked well for me. I have 1 question though in the official cudnn installation guide from NVIDIA it says to copy 3 files where you only say to copy 1. I was wondering why this is?

3. Copy the following files into the CUDA Toolkit directory:
Copy to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0\bin.
Copy <installpath>\cuda\ include\cudnn.h to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0\include.
Copy to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0\lib\x64.

[mariamal-ansary]

Thank you very much! this was extremely helpful!

[ManjunathMC]

Simple n nice, keep making videos buddy...

[ichbins]

No mention of Visual Studio which gives me the greates problems. Which boxes do I need to check so CUDA can successfully install?

[npoddubny]

Sorry for necro. May be you can help, I need tensorflow-gpu 1.8 but now it's only latest v2.2.0 is available. I have neural GAN that uses tf.contrib which is not supported. Thanks

[chaimaakabdi]

when i try to install tensorflow gpu they gave me that
no version found that satisfies the requirement

[JU-ymuf]

Thanks, I followed this video. very well.

[JonathanJMusic]

What if my GPU only has cuda capability of 2? I have a Nvidia Quadro 4000 GPU

[jojo-pqjn]

I have Geforce 1050ti on win 7 pc while running the 'hello world' command my output is below. I think its not fully detecting the gpu please correct me if am wrong

(tf15) C:\Users\admin>python
Python 3.6.5 |Anaconda, Inc.| (default, Mar 29 2018, 13:32:41) [MSC v.1900 64 bi
t (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
>>> import tensorflow as tf
>>> tf.__version__
'1.5.0'
>>> hello=tf.constant('hello tensorflow')
>>> with tf.Session() as sess:
... print(sess.run(hello))
...
2018-05-16 15:37:53.355099: I
Found device 0 with pro
perties:
name: GeForce GTX 1050 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.455
pciBusID: 0000:02:00.0
totalMemory: 4.00GiB freeMemory: 3.73GiB
2018-05-16 15:37:53.356099: I
Creating TensorFlow dev
ice (/device:GPU:0) -> (device: 0, name: GeForce GTX 1050 Ti, pci bus id: 0000:0
2:00.0, compute capability: 6.1)
b'hello tensorflow'
>>>

[era_ali]

Thank you I managed to install cuda and tensorflow gpu as you showed in this video. Now I am facing another problem while running darkflow. Though I have installed opencv but opencv runs in tensorflow cpu not in gpu. So can you tell me how I can install opencv for tensorflow gpu.

[ap-qnhw]

Can someone give me the result of "conda list" and a list of env variable linked to nvidia for cuda 9.0 ? It's not working for me. Thanks.

[lillebron]

I keep getting this error even after following this video exactly:
>>> import tensorflow as tf
Traceback (most recent call last):
File "C:\Users\nbala\Anaconda3\envs\tf18\lib\site-packages\tensorflow\python\platform\self_check.py", line 62, in preload_check

File "C:\Users\nbala\Anaconda3\envs\tf18\lib\ctypes\__init__.py", line 348, in __init__
self._handle = _dlopen(self._name, mode)
OSError: [WinError 193] %1 is not a valid Win32 application

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\Users\nbala\Anaconda3\envs\tf18\lib\site-packages\tensorflow\__init__.py", line 22, in <module>
from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
File "C:\Users\nbala\Anaconda3\envs\tf18\lib\site-packages\tensorflow\python\__init__.py", line 49, in <module>
from tensorflow.python import pywrap_tensorflow
File "C:\Users\nbala\Anaconda3\envs\tf18\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 30, in <module>
self_check.preload_check()
File "C:\Users\nbala\Anaconda3\envs\tf18\lib\site-packages\tensorflow\python\platform\self_check.py", line 70, in preload_check
% build_info.nvcuda_dll_name)
ImportError: Could not find 'nvcuda.dll'. TensorFlow requires that this DLL be installed in a directory that is named in your %PATH% environment variable. Typically it is installed in 'C:\Windows\System32'. If it is not present, ensure that you have a CUDA-capable GPU with the correct driver installed.


What could be the reason? And how do i fix it? Thanks

[pe]

tensorflow imported correctly but it is not detecting my gpu

>>> hello=tf.constant("hello world")
>>> with tf.Session() as sess:
... print(sess.run(hello))
…
and nothing is showing up

[ArjitGuptaAG]

I have the cpu version of tensorflow installed already. is there a way to install the gpu version over that or will i have to install it in a different environment. I have gtx 1050 ti.