Windows环境下在jupyter中安装tensorflow

1下载安装anaconda

首先，我们在anaconda官网上下载anaconda，具体的下载安装过程不再细说，在windows终端使用conda --version来验证是否安装成功：

C:\Users\hp>conda --version
conda 4.3.30

接下来使用jupyter notebook进入jupyter环境：

C:\Users\hp>jupyter notebook
[I 09:34:32.864 NotebookApp] JupyterLab alpha preview extension loaded from C:\Users\hp\Anaconda3\lib\site-packages\jupyterlab
JupyterLab v0.27.0
Known labextensions:
[I 09:34:32.864 NotebookApp] Running the core application with no additional extensions or settings
[I 09:34:33.400 NotebookApp] Serving notebooks from local directory: C:\Users\hp
[I 09:34:33.400 NotebookApp] 0 active kernels
[I 09:34:33.400 NotebookApp] The Jupyter Notebook is running at: http://localhost:8888/
[I 09:34:33.400 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).

在浏览器中输入http://localhost:8888/即可进入jupyter的开发环境。

2.配置tensorflow环境

在终端输入conda create -n tensorflow python=3.5，即可创建python版本为3.5的tensorflow环境，并使用activate tensorflow进行tensorflow环境：

C:\Users\hp>activate tensorflow

(tensorflow) C:\Users\hp>

并输入pip install tensorflow进行tensorflow的安装，鉴于安装时间过慢（且很大概率安装不成功），这里推荐使用清华的镜像进行下载：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple tensorflow

接下来分别使用conda install jupyter；conda install ipython；conda install spyder，在tensorflow环境中进行jupyter、ipython以及spyder的安装：

(tensorflow) C:\Users\hp>conda install jupyter
(tensorflow) C:\Users\hp>conda install ipython
(tensorflow) C:\Users\hp>conda install spyder

3.在tensorflow环境中进入jupyter

安装完成后退出windows终端，然后重新进入windows终端，使用activate tensorflow进入所创建的tensorflow环境，在tensorflow环境中输入jupyter notebook，即可进入可使用tensorflow的jupyter环境，每次进入jupyter环境的时候，重复这一段操作即可：（同理，可进入tensorflow环境下的ipython以及spyder）

C:\Users\hp>activate tensorflow

(tensorflow) C:\Users\hp>jupyter notebook
[I 10:02:49.472 NotebookApp] Serving notebooks from local directory: C:\Users\hp
[I 10:02:49.472 NotebookApp] 0 active kernels
[I 10:02:49.472 NotebookApp] The Jupyter Notebook is running at: http://localhost:8888/
[I 10:02:49.472 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).

#在jupyter中输入这段代码，运行成功则说明安装成功
import tensorflow as tf
print(tf.__version__)
1.6.0