When a machine learning model is released, it must be made available for downstream consumption; Verta offers endpoints as an API for configuring and releasing deployable models.
This tutorial will go through how to create and retrieve such an endpoint using Verta client and CLI.
Users can create an endpoint using
Client.create_endpoint() as follows:
endpoint = client.create_endpoint(path="/some-path")
This endpoint can be retrieved later on via its path or its ID using
endpoint = client.get_endpoint(path="/some-path")# alternatively:endpoint = client.get_endpoint(id=<endpoint id>)
We can achieve the same task using the endpoint command-line interface:
verta deployment create endpoint /some-pathverta deployment get endpoint /some-path