Vertex Model Garden models
With Vertex AI Model Garden, we are able to serve arbitrary models that are not available from a hosted endpoint. i.e., We are able to serve
Gemma on our own deployed endpoint while not being able to access it the same way we would access ClaudeOpus from Model Garden. In the Vertex AI plugin, we are able to reference ClaudeOpus as a value in the modelGarden configuratiuon options. Are we able to also reference these arbitrary endpoints we deploy via model garden or would some other plumbing need to take place?
