[machine vision learning] Anaconda download installation configuration and basic operation

Posted by digitalecartoons on Sat, 29 Jan 2022 23:29:38 +0100

001. Use Conda to manage Python environment in Anaconda

Tool: Anaconda Prompt (Anaconda) | please run it with administrator privileges

Update Anaconda

conda update conda

Create environment

conda create -n env_name package_names

env_name: environment name

package_names: package name, version can be specified

Case 1

Requirement: create evn_python2, python version is the latest version of 2, and pandas0.0 is installed at the same time Version 24.0, Numpy package

conda create -n evn_python2 python=2 numpy pandas=0.24.0

Switch environment


 conda activate my_env


source activate my_evn

Query the packages installed in the current environment

conda list

Leave the environment

conda deactivate

Environment sharing

Export environment

conda env export > environment.yaml

You can specify the path to save.

(python373) C:\Windows\system32>conda env export > d:\environment.yam

Import and update environment

(1) Enter the current environment first

conda activate Environment name

(2) Update environment

conda env update -f=/path/to/environment.yml

Where, - f indicates the local path of the environment file to be used; Set / path / to / environment Replace YML with the actual local path.

List environment

conda env list

The asterisk indicates the current environment

The default environment of Anaconda command line (that is, the environment used when no environment has been selected) is named base

Delete environment

conda env remove -n env_name

View environment information

conda info

002. Using conda to manage Python packages in Anaconda

Installation package

Search package

conda search requests

Installation package

conda install requests

In addition, if you can't find the package you want to install in the library provided by Anaconda, or you want to install an updated version, you can also call
Install through community maintained CONDA forge. For example, if you want to use CONDA forge to install Pandas, you can use
Give orders.

conda install -c conda-forge pandas

If the package you want to install cannot be found in the above libraries, you can also use the standard Python package management command pip to complete the installation in the current ring
Installation of third-party packages in the environment. For example, the tushare package used to obtain domestic financial and stock data can be used in the current ring
Use the following commands to complete the installation:

pip install tushare

Uninstall package and upgrade package

Uninstall package

conda uninstall packages_name

If you want to delete a package in another environment in the current environment, you can use the following command.

conda uninstall my_env packages_name

Upgrade package

Upgrade a package of the current environment

conda update my_env packages_name

Upgrade a package from another environment

conda update my_env packages_name

Upgrade all packages of the current environment

conda update --all

Add new libraries or remove libraries for Anaconda

add to

Using the mirror image of China University of science and technology

conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/

Displays the current mirror addresses

conda config --show-sources

The configuration information of conda is stored in In condarc file

  • windows system condarc is located in C:\Users \ username \ directory
  • Linux system/ condarc is located in / home / username / directory

Set the channel address (mirror source address) to be displayed during search

conda config --set show_channel_urls yes


conda config --remove channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/

Use conda config --show to confirm that the address has been removed.

conda config --show
conda config --show-sources

003.Anaconda basic usage exercise

1. Configure alicloud source OR Tsinghua source

conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
conda config --show-sources
conda config --set show_channel_urls yes
conda config --add channels https://mirrors.aliyun.com/anaconda/pkgs/free/
conda config --add channels https://mirrors.aliyun.com/anaconda/pkgs/main/
conda config --show-sources
conda config --set show_channel_urls yes

Delete source

conda config --remove channels URL

2. Check the existing environment and delete the useless environment

conda env list

conda env remove -n evn
conda env remove -n evn_python2

3. Configure the environment as required

Environment 1:

Python3.7.3 Tensorflow-gpu 1.13.1 Numpy1.16.3

conda create -n env_pythonA python=3.7.3 Tensorflow-gpu=1.13.1 Numpy=1.16.3

Environment 2:

Python3.10.0 TensorFlow-gpu 2.8.0 Numpy1.22.1

conda create -n env_pythonB python=3.10.0 Tensorflow-gpu=2.8.0 Numpy=1.22.1

An error occurred

Search TensorFlow package

The maximum version found is 2.6.0

conda create -n env_pythonB python=3.10.0 Tensorflow-gpu=2.6 Numpy=1.22

Incompatibility found between versions

According to the prompt,

When numpy=1.22.1, the python version is > = 3.8, ❤️. 9.0a0

When tensorflow GPU = 2.6.0, the python version is required to be 3.7 Or 3.8 Or 3.9*

conda create -n env_pythonB python=3 tensorflow-gpu=2.6.0 Numpy=1.22.1

At present, python version 3.9.7 is automatically selected.

(the process is a little long. You can have a cup of coffee)

4. Test in pychar

(1) Find environment path

conda info --env


conda env list

(2) Add environment to PyCharm


Build TensorFlow (google.cn) from source code in Windows Environment

004.Anaconda installs OpenCV

1. Switch environment

conda activate env_pythonA

2. Find the package about opencv

conda search opencv

3. Install opencv

conda install opencv

4. Import in pycharm

# Import opencv Library
import cv2 as cv

Topics: Python OpenCV machine vision