Vertex AI Metadata
Vertex AI Metadata is a repository of metadata that is generated across a variety of Vertex AI components. When models are developed in the ML workflow pipelines, metadata is generated and stored, and you can consolidate this metadata into a single metadata store, which allows users to query and answer questions such as the following:
- Which version of a trained model has achieved a certain quality threshold?
- Which pipeline runs/uses a certain dataset?
Within a Vertex AI pipeline, users can also configure data objects that are written to the metadata store. And from there, users can create Context objects that organize these data objects into logical groupings and obtain more insights.
Using the Vertex AI metadata API, users can build schemas, organize data objects, or query data that’s been stored within Vertex AI metadata.
