CentOS Anaconda(python3.6)安装tensorflow

Posted 2020-10-16 zhenzhen78

本来安装tensorflow是一件无比简单的事，但在我的电脑上却装了一个星期。期间遇到各种麻烦事、各种坑，在此记录一下，方便大家。报错包括：

• undefined symbol: zgelsd_
• ImportError: cannot import name ‘multiarray’
• whl is not a supported wheel

1，安装Anaconda

下载地址：https://www.continuum.io/downloads/（我安装的是linux-64-python3.6) 
我一开始是直接在python上装，但是python3.4(和python3.5）的numpy版本（1.12.0）似乎有问题，tensorflow可以安装成功，但是运行时调用numpy便报错了。报错如下：

import numpy Traceback (most recent call last):

File "", line 1, in

File "/usr/local/lib/python3.4/dist-packages/numpy/init.py", line 142, in from . import add_newdocs

File "/usr/local/lib/python3.4/dist-packages/numpy/add_newdocs.py", line 13, in from numpy.lib import add_newdoc

File "/usr/local/lib/python3.4/dist-packages/numpy/lib/init.py", line 18, in from .polynomial import *

File "/usr/local/lib/python3.4/dist-packages/numpy/lib/polynomial.py", line 20, in from numpy.linalg import eigvals, lstsq, inv

File "/usr/local/lib/python3.4/dist-packages/numpy/linalg/init.py", line 51, in from .linalg import *

File "/usr/local/lib/python3.4/dist-packages/numpy/linalg/linalg.py", line 29, in from numpy.linalg import lapack_lite, _umath_linalg

ImportError: /usr/local/lib/python3.4/dist-packages/numpy/linalg/lapack_lite.cpython-34m.so: undefined symbol: zgelsd_

在github https://github.com/numpy/numpy/issues/8697上也提问了，但是也没有解决我的问题 ：undefined symbol: zgelsd_。（期间还出现过ImportError: cannot import name ‘multiarray’ 这种问题，对于linux菜鸟完全不知道怎么办） 
这是numpy的问题，与tensorflow无关，但是我也迟迟无法解决。无果，转向直接安装anaconda，装好之后，numpy可以正常运行，tensorflow的安装却无比曲折。

2，安装tensorflow(cpu版）

对anaconda命令的熟悉，可以参考http://www.jianshu.com/p/d2e15200ee9b 
官方的建议是即时你有gpu，但也可以先装一个cpu版，创建环境的命令为：

conda create -n tensorflow python=3.6 
（一定要指定python版本，我一开始没有写python=3.6，后面各种失败）

先下载安装包，下载路径为：https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.0.0-cp36-cp36m-linux_x86_64.whl 
下载之后，将whl文件重命名为tensorflow-1.0.0-py3-none-linux_x86_64.whl，否则会出现

tensorflow-1.0.0-cp36-cp36m-linux_x86_64.whl is not a supported wheel on this platform.

一样的报错，具体参考https://github.com/tensorflow/tensorflow/issues/1990 
然后进入环境并安装tensorflow

source activate tensorflow #激活tensorflow环境
cd /Downloads #切换到whl文件所在文件夹
pip install --ignore-installed --upgrade tensorflow-1.0.0-py3-none-linux_x86_64.whl #切记，不要用sudo pip，也不要用pip3，然后--ignore-installed --upgrade等参数也不能省略，否则会出错。

 

3，安装tensorflow(gpu版）

创建环境的命令为：conda create -n tensorflow-gpu python=3.6 
先下载安装包，下载路径为：https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-1.0.0-cp36-cp36m-linux_x86_64.whl 
然后进入环境并安装tensorflow-gpu

source activate tensorflow-gpu #激活tensorflow环境
cd /Downloads #切换到whl文件所在文件夹
pip install --ignore-installed --upgrade tensorflow_gpu-1.0.0-cp36-cp36m-linux_x86_64.whl #切记，不要用sudo pip，也不要用pip3，然后--ignore-installed --upgrade等参数也不能省略，否则会出错。

接着，还要配置cuda和cudnn，可以到nvidia官网下载，接下来的配置可参考http://blog.csdn.net/jteng/article/details/52975247

4，验证安装

成功。

(tensorflow)$ python
import tensorflow as tf
hello = tf.constant(‘Hello, TensorFlow!‘)
sess = tf.Session()
sess.run(hello)

 

存在的问题，运行时，两个版本均有warning, NOT error，但是不影响结果，只是执行速度比较慢，据说是因为为了不同框架上的可迁移性，还没有对cpu进行编译，他建议你为了更快的速度，可以从编码编译，执行速度会更快。参考https://github.com/tensorflow/tensorflow/issues/8037

W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn‘t compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.

转自：http://blog.csdn.net/michaelliang12/article/details/60106686