Problem

You’re trying to create a Vector Search index. Whether you use the UI or the API, you receive an error.

Using the UI

While trying to create a Vector Search index using the UI, you receive the following error message. 

Index creation failed: Failed to call Model Serving endpoint <endpoint-name>

Using the API 

While trying to create a Vector Search index using the Databricks API, you receive the following error message in response to your call. 

"error_code":"INVALID_PARAMETER_VALUE","message":"Failed to call Model Serving endpoint: <endpoint-name>.","details":[{"@type":"type.googleapis.com/google.rpc.RequestInfo","request_id":"<request-id>","serving_data":""}]}

Cause

You are not using a text embedding model. When creating a Vector Search model, only text embedding models are permitted.

Solution

Use a text embedding model such as GTE Large (En) or BGE Large (En). 

For details, refer to the “GTE Large (En)” and “BGE Large (En)” sections of the Supported models for Databricks Foundation Models APIs (AWS | Azure | GCP) documentation.