Anaconda installation and Python environment creation under Linux offline state

1 Download and Installation Instructions

• download

  • Download address: https://repo.anaconda.com/archive/
  • Version: Here, the anaconda version 2020.11 is used as an example, and the python version it carries is 3.8.5.
  • Download: Find and download the Anaconda3-2020.11-Linux-x86_64.sh file in the above link, or enter https://repo.anaconda.com/archive/Anaconda3-2020.11-Linux-x86_64.sh to directly download the installation package

• File upload
  Upload the downloaded installation package to the file folder in the offline linux directory, upload it to the files folder here.

2 Anaconda installation

• Copy the cuda toolkit installer in the files folder to the software folder

  cp files/Anaconda3-2020.11-Linux-x86_64.sh software/
  

  Enter the directory where the cuda toolkit installer is placed (assuming the current directory is the user's home directory)

  cd software/
  # 查看当前目录
  pwd
  # 输出内容为：/data/users/CHDHPC/2017901437/software/
  

  Execute ls to view the files in the current directory:
  

• Modify the running permissions of the Anaconda toolkit installer

  chmod +x Anaconda3-2020.11-Linux-x86_64.sh
  

• Run the Anaconda toolkit installer

  ./Anaconda3-2020.11-Linux-x86_64.sh
  

• Press the Enter key Enter, the following content appears
  

• Continue to press Enter until the following content appears
  

• Type yes and press Enter. The following content appears
  

• At this time, you need to select the installation path of anaconda

  • option 1 : Press Enter to install it directly to the path "/data/users/CHDHPC/2017901437/anaconda3";

  • Option 2 : Enter the absolute path of the location to be installed, such as installing to the software directory under the home directory, you can fill in: "/data/users/CHDHPC/2017901437/software/anaconda3"
    

• Press Enter to start the installation. The following content appears to indicate that the installation is complete
  

• Then choose whether to initialize conda and add its environment variables in bashrc

  • Option 1 : Select "Yes", enter yes, press Enter, and the conda initialization will be completed automatically. The bashrc file will automatically add the following environment variables:
    

  • option 2 : Select "No", enter no, and press Enter to skip initialization. At this time, you need to manually complete the addition of environment variables:

    • The bashrc file is located in the user's home directory, namely "/data/users/CHHDHPC/2017901437/", so use the cd command to go to this directory

    • Execute vim .bashrcand add the following to it:

      export PATH="/data/users/CHDHPC/2017901437/software/anaconda3/bin:$PATH"
      

    • Execute wqsave and exit

    • Or if you choose no and still want to initialize, you can complete the initialization with the following command:

      source /data/users/CHDHPC/2017901437/software/anaconda3/bin/activate
      conda init
      

• Execute source ~/.bashrcto make the environment variables take effect; or disconnect the remote connection and re-log in to Gaosu remotely to make the environment variables take effect.

• The installation is complete
  

2.3 Gaoshu creates a virtual environment offline

• Configure conda (skip this step if you use the conda default path)

  • Execute. conda configA .condarc file will appear in the ~/ directory, which is used to configure personal conda

  • Enter the following configuration content:

    # 断网设备 .condarc
    show_channel_urls: true
    envs_dirs:
      - /data/users/CHDHPC/2017901437/software/myconda/envs
    pkgs_dirs:
      - /data/users/CHDHPC/2017901437/software/myconda/pkgs
    

    That is, specify the environment directory and package directory

• Download and upload packages (based on another networked linux device)

  • Execute. conda configA .condarc file will appear in the ~/ directory, which is used to configure personal conda

  • Enter the following configuration:

    # 联网设备 .condarc
    show_channel_urls: true
    envs_dirs:
      - /data/myconda/envs
    pkgs_dirs:
      - /data/myconda/pkgs
    

  • download package

    conda create -n 创建的环境名称 python==版本 --download-only
    # --download-only   表示只下载，不创建环境
    

  • upload package

    • When packaging networked devices pkgs, you must ensure that the package is complete. There may be a certain package in the default environment of the networked device, resulting in no repeated downloads. You need to manually pick it out. The package directory is/data/myconda/pkgs
    • Upload the packaged pkgs package to the user's home directory of Gaosuan using remote software
    • Copy the files in the uploaded pkgs to the pkgs directory of the home directory, the output path is/data/users/CHDHPC/2017901437/software/myconda/pkgs

• Download and upload packages (based on another networked Windows device)

  • download package

    conda create -n 创建的环境名称 python==版本 --download-only
    # --download-only   表示只下载，不创建环境
    

  • upload package

    • Same as above, pack the networked device pkgs, and then upload the packed pkgs package to the Gaosuan user’s home directory using remote software
    • Copy the files in the uploaded pkgs to the pkgs directory of the home directory, the output path is/data/users/CHDHPC/2017901437/software/myconda/pkgs

• Create a new virtual environment in Gaoshu anaconda

  conda create -n 创建的环境名称 python==版本 --offline
  

• activate and verify

  • activation

    # 激活(是否需要看情况)
    source activate
    # 切换环境
    conda activate 创建的环境名称
    

  • verify

    # 查看python版本
    python --version