In this lab, you deploy two models on Cloud AI Platform, and then use the What-If Tool to identify potential bias in the models and underlying dataset. You will have the opportunity to address identified bias, retrain the models, redeploy them, and continue to compare them with the What-If Tool.

Objectives

In this lab, you will perform the following tasks:

• Train tf.keras and Scikit Learn models in Cloud AI Platform Notebooks.
• Deploy models to AI Platform Predictions.
• Use the What-If Tool to analyze and compare the models.

Start a JupyterLab Notebook instance

1. In the GCP Console, in the Search bar, type in Notebook.
2. Select Notebooks for AI Platform.

https://cdn.qwiklabs.com/8Kn7H9l27eMm%2BB6emya%2FKViTYgpJatwo1uBp6vIDUGQ%3D

3. On the Notebook instances page, click New Instance.

4. In the Customize instance menu, select the latest version of TensorFlow without GPUs.

https://cdn.qwiklabs.com/z67oBeoigDBwLvDhGe616v8Z8z0biDhF9R6VklELUT0%3D

5. In the New notebook instance dialog, for Region, select us-central1, for Zone, select a zone within the selected region, leave all other fields with their default options, and click Create. .

After a few minutes, the AI Platform console will display your instance name, followed by Open Jupyterlab.

6. Click Open JupyterLab. Your notebook is now set up.

Clone the sample code

To clone the training-data-analyst notebook in your JupyterLab instance: