ananconda安装机器学习深度学习框架

安装pytorch:

conda install pytorch-cpu torchvision-cpu -c pytorch

安装opencv：

直接使用conda intall opencv 系统默认3.4.2

注：在python文件中import cv2出现无法识别的现象，见下面

安装keras与tensorflow：

conda activate base

conda install keras

conda install tensorflow

安装Caffe与MXNet深度学习框架：

conda install Caffe

conda install mxnet

解决anaconda中已经安装了opencv3，但无法import的问题

你可能遇见的问题：

1. ImportError: No module named cv2
2. ImportError: libz-a147dcb0.so.1.2.3: cannot open shared object file: No such file or directory

第一个问题：

你可能没有成功安装opencv环境，或者你需要重新安装opencv的环境进行尝试

你可以查看这一篇博客进行配置opencv卸载与安装

第二个问题：

你可能查看了某些博客，要求你将cv2.so移动到你的使用的python中（ sudo cp /usr/local/lib/python2.7/dist-packages/cv2/cv2.so /home/username/anaconda3/lib/python3.6/dist-packages/）,然后你再次import cv2就会发现会出现以上的报错

你可以通过which python命令查看你的python的来源。例如你通过which python得到的结果是/home/learner/anaconda3/bin/python

那么说明你需要的opencv文件应该放在anaconda3/lib/python3.6(你的对应的版本)/dist-packages/中

解决方法:

如果你确实已经按照以上的博客进行安装好opencv了，那么你可以继续进行，否则，请先按照以上博客进行opencv的安装。

根据提示，缺少的是libz-a147dcb0.so.1.2.3这个文件，那么我们进行查找通过sudo find / -name libz-a147dcb0.so.1.2.3

得到结果如下：

find: ‘/run/user/1000/gvfs’: Permission denied
/var/lib/docker/overlay2/f72d8ff7ad2486a45a06a7bccdebe2628377a839c7f943d7e2b06f44a825be87/diff/usr/local/lib/python3.6/dist-packages/PIL/.libs/libz-a147dcb0.so.1.2.3
/var/lib/docker/overlay2/f72d8ff7ad2486a45a06a7bccdebe2628377a839c7f943d7e2b06f44a825be87/diff/usr/local/lib/python3.6/dist-packages/h5py/.libs/libz-a147dcb0.so.1.2.3
/var/lib/docker/overlay2/f72d8ff7ad2486a45a06a7bccdebe2628377a839c7f943d7e2b06f44a825be87/diff/usr/local/lib/python3.6/dist-packages/matplotlib/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tensorflow/lib/python3.6/site-packages/PIL/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tensorflow/lib/python3.6/site-packages/h5py/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tensorflow/lib/python3.6/site-packages/matplotlib/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tensorflow/lib/python3.6/site-packages/cv2/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tf/lib/python3.6/site-packages/PIL/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tf/lib/python3.6/site-packages/h5py/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tf/lib/python3.6/site-packages/matplotlib/.libs/libz-a147dcb0.so.1.2.3
/home/learner/.conda/envs/tf/lib/python3.6/site-packages/cv2/.libs/libz-a147dcb0.so.1.2.3
/usr/local/lib/python3.5/dist-packages/cv2/.libs/libz-a147dcb0.so.1.2.3
/usr/local/lib/python3.5/dist-packages/matplotlib/.libs/libz-a147dcb0.so.1.2.3
/usr/local/lib/python2.7/dist-packages/h5py/.libs/libz-a147dcb0.so.1.2.3
/usr/local/lib/python2.7/dist-packages/cv2/.libs/libz-a147dcb0.so.1.2.3
/usr/local/lib/python2.7/dist-packages/matplotlib/.libs/libz-a147dcb0.so.1.2.3

我们可以观察到有很多cv2的文件夹，而我们python所使用的anaconda3/lib/python3.6/site-packages中没有这个文件，那么将cv2.so文件复制过去

cp /usr/local/lib/python2.7/dist-packages/cv2.so /anaconda3/lib/python3.6/site-packages

在此处还不够，需要加上以下声明，可以添加到.bashrc中

export PYTHONPATH=/home/learner/anaconda3/lib/python3.6/site-packages/:$PYTHONPATH

验证结果

Python 3.6.3 |Anaconda, Inc.| (default, Oct 13 2017, 12:02:49) 
[GCC 7.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>>