My computer did not install a dual system, so decided to install a tensorflow in the virtual machine, the following is the installation process:
1. Installing Anaconda2 for Linux
The official website under the words very slowly, to the Tsinghua Mirror website, I last article has the website
Installation: Bash anaconda2.sh
Next, you can choose whether to create a virtual environment, create the words Conda create-n tensorflow python=2.7 to create a virtual environment named TensorFlow
After creation you can follow the prompts source activate TensorFlow enable the virtual environment, source deactivate shutdown environment
Then you can enter the two different environments at the same time the Conda list contains the module, create a virtual environment for the reasons for you to distinguish, the province of the original Python environment to make a mess,
That is, convenient management, to remove the words Conda remove-n tensorflow--all, specific commands Conda official documents
2, Installation TensorFlow
After activating the TensorFlow environment, enter the following command:
Pip install--ignore-installed--upgrade https://storage. Googleapis. com/tensorflow/linux/cpu/tensorflow- 0.8. 0rc0-cp27-none-linux_x86_64. WHL
You will see the modules to be downloaded online, some very fast, some very slow, I download numpy is very slow, and then timeout, in a few retries or unsuccessful, so I was wondering whether to turn over a wave of wall
With the blue light after the FQ or useless, but also in the virtual machine in the network problem, I see others are not installed in the virtual machine does not appear this problem, instead of net mode sharing network card or not,
So on-line installation, the need to download a few WHL files are first downloaded, there are numpy,six,protobuf,wheel,setuptools, the corresponding version of the requirements, specific to see you install TensorFlow version,
finally in order each PIP install is good, the last pip install TensorFlow, if Import TensorFlow Yes, it's all set up.
Reference: http://blog.csdn.net/tina_ttl/article/details/51762471
Ubuntu14.04 under TensorFlow installation