Vertex AI on Google Distributed Cloud (GDC) air-gapped includes three APIs, one for each of the following pre-trained models:
- Optical Character Recognition (OCR): To learn more about OCR on GDC, see Detect text in images.
- Speech-to-Text: To learn more about Speech-to-Text on GDC, see Transcribe audio.
- Translation: To learn more about Translation on GDC, see Translate a language into English.
Vertex AI on GDC also includes Vertex AI Workbench, which you can use to create a JupyterLab notebook.
Before you begin
Use the GDC console to enable, disable, and view the endpoints of the Vertex AI services. Before you can enable the pre-trained APIs, you must do the following:
Sign in to the GDC console. If you aren't able to sign in, see Connect to an identity provider. If you must sign out, see Sign out.
Create and grant permissions to users allowing them to enable pre-trained APIs.
Provision the GDC system cluster with GPUs. For more information, contact your Infrastructure Operator (IO) and see Enable GPU support for user clusters.
Configure the GDC ingress gateway. For more information, see Project network policy.
Set up the GDC domain name system (DNS). If you haven't set up the DNS, work with your Infrastructure Operator (IO) to complete this prerequisite.
Permissions required to enable pre-trained APIs
To get the permissions you need to enable the pre-trained APIs, ask your
Organization IAM Admin to grant you the AI Platform Admin (ai-platform-admin)
role in your project namespace. If you don't have the correct permissions, the
GDC console doesn't display the buttons to enable or disable the
pre-trained APIs.
Only the Organization IAM Admin can grant the AI Platform Admin
(ai-platform-admin) role to users.
Enable pre-trained APIs
You can enable the OCR, Speech-to-Text, and Translation pre-trained APIs using the GDC console.
To enable the pre-trained APIs, do the following:
If you aren't signed in to the GDC console, follow the steps in Sign in to the GDC console.
Verify that the prerequisites required to enable pre-trained APIs are complete. You might need to work with your Infrastructure Operator (IO) to complete the prerequisites.
In the navigation menu, select Vertex AI, and click Pre-trained APIs.
Click Enable All APIs or Enable a specific service. If the buttons aren't visible, verify that you have the correct permissions to enable the pre-trained APIs. To learn more, see Permissions required to enable pre-trained APIs.
In the confirmation dialog, click Enable. A progress message displays. The enablement duration varies. It might take 15-45 minutes depending on the state of the System Cluster.
If you want to view the status of the pre-trained APIs, see View service statuses and endpoints.
If the services are taking too long to enable, the VAI-A0001 alert
(Enabling State Time Limit Reached) might trigger. To address this alert, your
IO should follow the steps described in the VAI R0001 runbook documentation.
Disable pre-trained APIs
You can disable the OCR, Speech-to-Text, and Translation pre-trained APIs using the GDC console.
To disable the pre-trained APIs, do the following:
If you aren't signed in to the GDC console, follow the steps in Sign in to the GDC console.
In the navigation menu, select Vertex AI, and click Pre-trained APIs.
On the Pre-trained APIs page, click Disable All APIs or Disable a specific service. If the buttons aren't visible, verify that you have the correct permissions to disable the pre-trained APIs. To learn more, see Permissions required to disable pre-trained APIs.
In the Disable pre-trained APIs confirmation dialog, enter disable to confirm that you want to take that action and then click Disable. A progress message displays.
If you want to view the status of the pre-trained APIs, see View service statuses and endpoints.