One: system and environment

System: Windows10 GPU: NVIDIA GeForce GTX TITAN X must first determine the version of Tensorflow that you want to install and the various versions of Tensorflow that you want to install The first step is to determine which version of Tensorflow you want to install and its corresponding versions

Determine the version

  • MSVC stands for Visual Stutio

Two: Installation steps

1. Install Visual Stutio 2017

Visual StudioYou can select the version here, and then install the Community.

2. Install CUDA 10.0

Then install all the way:Copy the code

Note: Use the default path!Command line to test whether the installation is successful: NVCC -v

3. Install cuDNN 7.4

3.1 cuDNN Select the corresponding version from Archive. After downloading, press the folder and copy the bin/,include/,lib/ folders to the CUDA installation directory directly. C: Program Files NVIDIA GPU Computing Toolkit CUDA v10.0 3.2 Add the directory of CUDA to the environment variable.

C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0;
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib;
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include;
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib\x64;
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin;
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\libnvvp;
Copy the code

4. Create tF-GPU environment, install TensorFlow-GPU 1.14, use TSINGHUA University PIP source

> conda create -n tF-14 PIP python=3.7 # install tF-14 PIP install --ignore-installed --upgrade Tensorflow - gpu = = 1.14.0 -i https://pypi.tuna.tsinghua.edu.cn/simple # PIP install tf - the gpu versionCopy the code

5. Test it out

To enter Python mode, enter import tensorflow as tf! [image_1e1itil7k114u1cn9881kl91en515. PNG 29.7 kB] [8] ps: FutureWarning here is not in conformity with the numpy version, can be ignored. PIP uninstall numpy (till you uninstall all versions) PIP install numpy==1.16.4Copy the code