Install Tensorflow gpu version in Windows

-

NVIDIA

TensorFlow






How to install the Tensorflow_gpu in Windows 10:
  • ·       At first go to the Tensorflow Official website to download Tensorflow gpu version.        Go to -> https://www.tensorflow.org/install/install_windows . I usually prefer you to Install  it in theAnaconda platform as: https://www.anaconda.com/download/
  • To install TensorFlow gpu version, i suggest you to install it by create a new conda environment. Here, is the Instruction:           
  • conda create -n tensorflow-gpu python=3.5      # based on your version of python            
    
activate tensorflow-gpu
    

    
pip install tensorflow-gpu
  • ·  Then go to the nvidia website to download cuda toolkit v8.0 .Since tensorflow -gpu is currently  working fine with v8.0 at:  Click Here .Here you can download.
  • ·   When the installation is finished. Then, download another file cudnn-8.0-windows10 (32 or 64 bit) from :  https://developer.nvidia.com/rdp/cudnn-archive . This is the DLL file help to run the tensorflow-gpu in your Computer.


  • ·                 After that extract the cudnn-8.0-windows10 from if it is in zip format. Then, go to the bin         directory and see cudnn64_6.dll file just copy it into directory usually it is found at -                    C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin where the                  cudart64_80.dll is found. And copy it inside the                                                                C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin  Folder. 
·         If the path is not set on your windows environment variable, then you can set as:  
·         Name of Environmental Variable as: path (i.e usually you to do set up the path in the Advanced System Settings.
·         Path to the cudnn64_6.dil as    : C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin

Validate your installation


    Then check it from the command line  as: 
Invoke python from your shell as follows:








python











>>> import tensorflow as tf
>>> hello = tf.constant('Hello, TensorFlow!')
>>> sess = tf.Session()
>>> print(sess.run(hello))

If the system outputs the following, then you are ready to begin writing TensorFlow programs:







Hello, TensorFlow!
If you are new to TensorFlow, see Getting Started with TensorFlow.

If there is no error then, it indicates that you set tensorflow-gpu version sucessfully in your PC.  ALL THE BEST:  👍

If you want to install the CPU Version then, please see the instruction at : Install TensorFlow CPU

This is created by Sushant Gautam. If you have any query please ask at the Comment.

For suggestion and feedback please send me at: Sushant1234gautam@gmail.com

Any ideas and feedback are Acceptable! Thanks a lot for reading this. 😐

Please see my github account page : https://github.com/sushant097/Tensorflow-Tutorial for more details.


Location: Kathmandu 44600, Nepal

Comments

Post a Comment

Popular posts from this blog

Basic Tensorflow Tutorial

-
Image
Overview Of Basic Operation in  Open source software library TensorFlow: Introduction: TensorFlow is an open-source software library dataflow programming across a different task.It is math library where it is mostly used in machine learning such as neural network. Google made it and publish it in 2009. Nowadays, it is used in many machine learning application. Some of the application are: RankBrain SmartReply(Neural Machine Translator i.e Google Translate) Google Brain Self Driving Car and so on. Besides it, Due to it's efficient model structure, robustness and open source, it is one of the leading machine learning libary . As, Today due to increase of the demands of the people and Rapid development of the Technology traditional Programming learning approach is not sufficient to develop the system i.e Robustness, Automatic, User-Friendly and so on. So, To accomplish such task and features we need modern Machine Learning approach which is provide by Tensorflow.
Read more