Table of Contents of Series Articles
Chapter 1 Environment Configuration for Getting Started with TensorFlow Deep Learning
Table of contents
Table of Contents of Series Articles
Preface
Currently, Tensorflow is supported on Windows, Linux and MacOS. This article will take the installation of Windows system as an example.
Before installing Tensorflow, we need to install Anaconda first, because it integrates many Python third-party libraries and their dependencies, which is convenient for us to call directly in programming.
Without further ado, let’s take a look at how to configure the environment!
1. Install anaconda
First enter the official website of anaconda:
Anaconda | Anaconda Distribution
After entering the official website, just download the latest version of anaconda. After using it a few times, I found that the new version is easier to use.
If you cannot download it, you can enter the domestic mirror download:
After the download is complete, click Install.
2. Install Tensorflow
1 Create an environment
Open Anaconda Prompt and enter the following code to create an environment:
conda create -n tensorflow tensorflow python=3.6
tensorflow is the name of the created environment;
python=3.6 means that in the created environment, the installed python version is 3.6
After entering the code, press Enter and wait for the installation to be successful.
2 Install Tensorflow
Open Anaconda Prompt and enter the following code to enter the environment you just created:
conda activate tensorflow
After entering the tensorflow environment, we can start installing tensorflow
Enter the code and install tensorflow:
python -m pip install tensorflow
If the installation process is too slow, you can use a mirror installation:
python -m pip install tensorflow -i https://pypi.douban.com/simple
Just enter the code and press Enter, and wait for the installation to be successful.
3 Install jupyter notebook
There is no jupyter in this environment, so we need to install it manually, otherwise it will be wrong to open the jupyter environment.
Enter the following code to install jupyter notebook:
pip install jupyter -i https://pypi.douban.com/simple
In order to check whether the installation is successful, you can enter:
conda list
Check whether the package is installed successfully