Additional permissions – Pipelines using Kubeflow for Custom Models

We also need to grant additional permission to the service account associated with the compute engine of GCP since we are fetching data from BigQuery. Follow these steps to grant the required permission.

Step 1: Open IAM and admin section

Follow the steps mentioned in Figure 7.2 to add roles to the service account associated with compute engine of GCP:

Figure 7.2: IAM of GCP

1. Click IAM.
2. Choose the service account associated with the project.
3. Click Edit Principal.

Step 2: Additional role

Follow the steps mentioned in Figure 7.3, 7.4 and 7.5 to edit the roles of the service account of compute engine:

Figure 7.3: Edit roles of service account

1. Click + Add Another Role and follow the steps mentioned in the Figure 7.4:

Figure 7.4: Adding BigQuery admin role

1. Select BigQuery admin role under the BigQuery product/service.
2. Click SAVE. A new role will be added to the service account as shown in the following figure:

Figure 7.5: New role added to service account is visible in the IAM

1. New role will be visible for the service account associated with the compute engine.

Creation of Workbench

In all the previous chapters we have worked on the workbench. Vertex AI Workbench and custom model training guides on creating a workbench using the GUI approach. Follow these steps to create Vertex AI workbench from the cloud shell:

Step 1: Activate cloud shell

Follow the steps shown in Figure 7.6 to activate the cloud shell and type the following commands:

Figure 7.6: Activate cloud shell

1. Click Cloud shell to activate it
2. Type in the following mentioned commands in the cloud shell to create workbench:

gcloud notebooks instances create kubeflow-custom-ml \

     –vm-image-project=deeplearning-platform-release \

     –vm-image-family=common-cpu-notebooks \

     –machine-type=n1-standard-4 \

     –location=us-central1-a

Step 2: Created workbench

The created workbench will be listed under workbench section of Vertex AI as shown in Figure 7.7:

Figure 7.7: Workbench creation complete

1. Newly created workbench.
2. Click on OPEN JUPYTERLAB.