Generate Embeddings

In these examples, you can see how to use the VECTOR_EMBEDDING SQL function or the UTL_TO_EMBEDDING PL/SQL function to generate a vector embedding from input text strings and images.

• Convert Text String to Embedding Within Oracle AI Database Perform a text-to-embedding transformation by accessing a vector embedding model stored in the database.

• Convert Text String to BINARY Embedding Outside Oracle AI Database Perform a text-to-BINARY-embedding transformation by accessing a third-party BINARY vector embedding model.

• Convert Text String to Embedding Using Public REST Providers Perform a text-to-embedding transformation, using publicly hosted third-party embedding models by Cohere, Generative AI, Google AI, Hugging Face, OpenAI, or Vertex AI.

• Convert Text String to Embedding Using the Local REST Provider Ollama Perform a text-to-embedding transformation by accessing open embedding models, using the local host REST endpoint provider Ollama.

• Convert Image to Embedding Using Public REST Providers Perform an image-to-embedding transformation by making a REST call to the third-party service provider, Vertex AI. In this example, you can see how to vectorize both image and text inputs using a multimodal embedding model and then query a vector space containing vectors from both content types.

• Generate Multi-modal Embeddings Using CLIP This section provides end-to-end instructions from installing the OML4Py client to generating multi-modal embeddings using CLIP.

• Vectorize Relational Tables Using OML Feature Extraction Algorithms