The first problem facing learning tensorflow is that the environment is configured, various python versions, tensorflow versions and various application packages
Now solve all problems with one software: anaconda
It can do it: a brand new computer, nothing is installed, just install the anaconda software and you can start development
- download
https://www.anaconda.com/products/individual#download-section
After installation, open the terminal command tool that comes with the mac, enter jupyter notebook and wait a while
If you need to change the default storage path of files, you can close jupyter and enter in the terminal
jupyter notebook --generate-configThe system will automatically generate a file and return the file name and address
Writing default config to: /Users/jiangshanruciduojiao/.jupyter/jupyter_notebook_config.pyThen copy the above path, and then edit the file
vim /Users/jiangshanruciduojiao/.jupyter/jupyter_notebook_config.pyThen enter the diagonal bar to search notebook_dir
/notebook_dirRemove the hash mark in front and add the full path of the directory you want to store in the back
Then press esc, then enter colon ➕ wq to save
Then start again
If it is windows, remember to remove the path behind the jupyter startup shortcut
Anaconda comes with many commonly used python libraries by default, such as numpy, scipy, etc.
But it does not come with tensorflow, so you need to install it yourself
How to install it?
Open the anaconda software, open the command line as shown in the figure below, execute the following installation commands, the Ali source speed is still relatively fast
pip install tensorflow-cpu==2.2.0 -i https://mirrors.aliyun.com/pypi/simple/Successfully installed appears to indicate successful installation
If you want to install other libraries, you can use a similar method, don’t fill in if you don’t know the version number, and download the latest version by default
pip install xxxxxxx -i https://mirrors.aliyun.com/pypi/simple/