2020 March 6
Basic environment to build
- System installed in UBUNTU 16.04 anaconda3., The selected version is python3.6, so select Tsinghua mirror , proceeds to select the corresponding version downloaded . Get file Anaconda3-5.2.0-Linux-x86_64.sh, need to be installed with a bash. The method of mounting details and verification process . In the process of conda be UPDATE, the need to download a file, it is proposed to replace the mirror source, method explain .
Then we have to do is anaconda environment to create a virtual environment (cpuVersion) , and name it tensorflow. Creating later be viewed by conda info --envs, you can see tensorflow environment has been established out. Inside create a virtual environment, there is a problem that needs to statements
~$pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.7.0-cp36-cp36m-linux_x86_64.whlConfirmed the version of tensorflow good! 1.7.0 given here!
Is the usual waypip install tensorflow=1.12
tensorflow linux 1.12 whl package file can be found
https://download.csdn.net/download/weixin_41369892/10806403
Here you need to remember the path where the virtual environment created, / home / gs / anaconda3 / env3 / tensorflow, need to use later in the process pycharm project interpreter set in!
New in pycharm in engineering MNIST_Sun2020.
Select FILE-> Setting, select Project interpreter, Keyihuode default interpreter for the python2.7, this is ubuntu comes with the environment, needs to be changed! Using Exiting enviroment and choose
/home/gs/anaconda3/envs/tensorflow/bin/python as the basic interpreter!
Reading comprehension
Are reading literature: tensorflow cross-entropy of understanding .
