Clicking the [Create API] tab of a created prediction model takes you to this screen. For the API specifications, refer to “Inference API Specifications.”
Step 1. Clicking the [Create API] tab with the model that you want to create a prediction API for selected takes you to this screen.
Step 2. Click the [Launch] button
Step 3. Select the API type in the [Start API creation] window, then click the [Start] button.
Step 4. The building API screen will be displayed. It takes about 5 to 20 minutes to build the API. The prediction API can be used after building is complete.
The following screen is displayed when the API has been created.
Check [API information].
Check [API request count]. The number of requests is counted daily and displayed as a graph.
Check [API status].
Step 1. Click the [Configure IP whitelist] button
Step 2. In the text box of the displayed window, enter the whitelist of IP addresses to allow. Enter the addresses in the CIDR format.
Step 3. Click the [Add] button.
To delete an IP address from the IP whitelist, click the [Delete] button on the right of the corresponding IP address.Click the [Stop] button.
With the prediction API stopped, click the [Delete] button. When you delete an API, the number of API options in use decreases and you can create an API with another model.
Click the [Lock API] button. The API is locked and cannot be stopped. This prevents users from accidentally stopping the API.
With the prediction API locked, click the [Lock API] button. Only an account with the administrator role can perform this operation. The administrator role is configured in the User Portal.