When using open-source embeddings models with SentenceTransformer, the models are fetched and loaded into memory. When the class is instantiated, add an install command-line option to pre-install the supported models.
This is useful for pre-caching model data in, for example, Docker containers in preparation for future use.