Skip to content

Installing the MissingLink Package for Use With Your Deep Learning Framework

The MissingLink package includes the following components:

  • MissingLink SDK: a lightweight library that wraps your favorite deep learning framework and makes your experiment monitorable remotely.
  • MissingLink CLI: a command-line tool for advanced integration.
  • Web dashboard: a simple yet powerful interface for monitoring and managing your experiments.

Requirements

MissingLink supports the following versions of Python:

  • Python 2: Version 2.7 or later
  • Python 3: Version 3.4 or later

For more information, see Python Package Index.

Installation instructions

Note

You must install MissingLink in the same virtual environment as your deep learning framework.

Install the MissingLink package using either the pip or conda package manager:

Using pip

sudo easy_install pip  # install pip if you have not yet
pip install missinglink

Using conda

conda config --add channels conda-forge
conda config --add channels missinglink.ai
conda install missinglink

It is highly recommended you perform the installation in a virtual Python environment using virtualenv.

Authentication

Before accessing your AI projects using MissingLink, you need to first receive authentication.

In a terminal, run the following command:

ml auth init

A browser opens for you to log in as when you log in to your MissingLink admin dashboard. After a successful login, you will be notified of success, for example "User authenticated. Close this tab and go back to the console."

In case you are running ml auth on machines without browsers, it is recommended to use the --disable-webserver flag. Instead of opening a browser, the system will give you a URL and prompt you to enter a token. Using the URL on a different machine, with a browser will provide the token that you need to enter.