Add Google's embedding and caching to vertex ai

[Godrules500]

Feature Description

Being that Vertex AI is covered under GCP company agreements, I am unable to use google generative ai. Because of this, I'd like for the features that google has to be added to the Vertex ai provider.

Use Case

Being able to work with caching to speed up the use of repetitive documents or using it with an extensively complex and long set of system instructions.

The ability to work with embeddings datasets pulled in from an API and used for rag searching.

Additional context

No response

[lgrammel]

think this is supported now