Skip to content

Running Experiments with Explicit Project and User Identifiers

This topic shows you how to run an experiment using specific project token and owner ID.

Normally, when you run an experiment, MissingLink assumes you wish to run it on the most recently selected project and as the user that was validated by the most recent ml auth init command.

MissingLink allows you to run experiments from a different project and as a different user, by specifying them explicitly. For example, if you integrated MissingLink with Keras, when you first create the KerasCallback instance, instead of writing:

missinglink_callback = missinglink.KerasCallback()


missinglink_callback = missinglink.KerasCallback(owner_id='', project_token='')

To obtain the token and owner ID that are required for this function call, see below.

For TensorFlow

missinglink_project = missinglink.TensorFlowProject(owner_id='', project_token='')

For PyTorch

missinglink_project = missinglink.PyTorchProject(owner_id='', project_token='')

For PyCaffe

missinglink_callback = missinglink.PyCaffeCallback(owner_id='', project_token='')

For scikit-learn

project = missinglink.SkLearnProject(owner_id='', project_token='')

Obtaining the token and owner ID of a project

Perform the following procedures to obtain the token and owner ID of a project:

  1. Display the list of projects.

  2. Click the menu icon at the extreme right of the desired project and select Integration info.

    Click Integration Info to display project token and owner ID

    The project token and owner ID are displayed:

    Lets you copy project token and owner ID

  3. Click Copy to copy the data to the clipboard. Then paste it to your code.