Deployments

Overview

After training an AutoML model, you can create a Deployment to submit requests to the model via an API endpoint. A deployment is a version of your model that is ready for inference requests. A deployment points to a Hive Data project that lets you submit requests to the model in the same way that you would interact with any Hive pre-trained model.

Create a Deployment

To create a deployment, go to a detail page for the model you would like to deploy and click the Create Deployment button in the top right.

The `Create Deployment` button lies at the top right of the detail page for an individual model.

All deployments for a model can be seen under the deployments tab.

The `Deployments` tab on the details page for a model shows all deployments of that model.

Clicking the name of a deployment opens the corresponding Hive Data project, allowing you to access API keys and view information about all submitted inference requests.

To begin using your model, click on the API Keys button on the top right of the Hive Data project page to copy your API Key. For instructions on how to submit a task via API, either synchronously or asynchronously, see our API Reference documentation.

Update Deployment

If you'd like to move a deployment from one training to another, you can update your deployment. This option is useful in the case that you want to continue to use the same API key, but you want to switch version of the model you're using. To do this, open the Deployments tab and click on the Update button located to the right of the Undeploy button (see the image below). After clicking on the button, you'll just need to choose which model version you want to update to.

To update an individual deployment, just click the `Update` button.

Video Inference

If you would like to submit videos to an Image Classification model, you will need to check the Use Video Inference checkbox on your Create Deployment form. Deployments without this option will only be able to support images.