Downstream Models

Downstream Models are task-specific models built on top of a trained Foundation Model. They allow you to define a concrete prediction objective — such as classification (binary, multiclass, or multilabel), regression, or recommendation — and train a model tailored to that goal. Each Downstream Model is configured with a target function, a training audience, and a schedule.

This section describes how to navigate and configure Downstream Models.

There are 2 sections:

Downstream Models View - Manage your Downstream Models

Downstream Model Configuration - Build and configure a Downstream Model