Oct/30/2018 Updated by

Jupyter

Starting Jupyter Notebook Server


[Up] Japanese English

[Prerequisite Conditions]

Setting Token or Password to Access Jupyter Notebook Server

  1. Generating Setting file.
  2.   [on Command Prompt]
      
    jupyter notebook --generate-config
    Location of Setting file: nitta should be replaced with your username.
  3. [Option] Edit the setting file and set token. token is a plaintext.
  4. In the example, xxx is token. Use string that are hard to guess from others and easy to memorize on your own.

    c.NotebookApp.token='xxx'
  5. [Option] Edit the setting file, and set password. password is an encrypted string.
  6. The way to encrypt is as follows. Type your own password in xxx. 他人にわかりにくく、自分では覚えやすい文字列を使うこと。 Use string that are hared to guess from others and easy to memorize on your own. The character string displayed as sh1: is the encrypted password.

      [on Conda]
    
    python >>> from notebook.auth import passwd >>> print(passwd()) Enter passwd: xxx Verify passwd: xxx sh1: 123456789abcdefg...... >>> exit()

    Write the encrypted password as the value of c.NotebookApp.password. In the example below, 123456789abcdefg......' is the encrypted password.

    c.NotebookApp.password = '123456789abcdefg......'
    [Note] For the security, it is better to keep comments (with # at the beginning of the line) for the next parameter.
    # c.NotebookApp.allow_remote_access = False

Starting Jupyter Notebook Server

  1. Activate python environment with Conda. Here, we assume that you use the previously prepared deep environment.
    (base) C:\Users\nitta> conda activate deep (deep) C:\Users\nitta>
  2. Move to the folder which the data of jupyter notebook is stored.
  3. (deep) C:\Users\nitta> cd Documents (deep) C:\Users\nitta\Documents> cd jupyter
  4. Start jupyter notebook.
  5. (deep) C:\Users\nitta\Documents\jupyter> jupyter notebook

[Note] With GPU, depending on the program, it is about 30 times faster. Therefore, it is better to use gpu environmentif possible.a However, if multiple kernels are operated, GPU allocation may fail, and the program may terminate prematurely. Be careful not to run multiple kernels using GPU at the same time.