Skip to Main Content
Gimme a Vector, Victor: Deploying Oracle Database 23ai VECTOR Datatype for Retrieval-Augmented Generation (RAG)

About This Workshop

Youtube Video

About This Workshop
Retrieval-Augmented Generation (RAG), when enabled through Oracle Database 23ai’s new VECTOR datatype, makes it simple to build APEX and analytic applications that leverage Large Language Models (LLMs) to solve complex business problems, even when the end user doesn’t even know what specific questions to ask.
This lab demonstrates:
- The basics of VECTORs and how they’re useful for LLMs and RAG
- How to build an experimental environment for leveraging 23ai VECTORs
- How to leverage LLMs to generate meaningful sample data without manually scraping it from external sources
- What business use cases are likely to find LLMs and RAG useful (and those that won’t)

Workshop Info

1 hour, 30 minutes
  1. Initialize the LiveLab Environment
  2. Prepare Database Objects
  3. Import ONNX Trained GenAI LLM
  4. Build Document Corpus for LLM Vocabulary
  5. Embed Corpus Via 23ai VECTOR Datatype 
  6. Create Vector Indexes
  7. Create ChatBot for LLM Interaction
  8. Deploy RAG Technology to Interact with ChatBot

Please note: In order to complete lab 7 & 8, you need to have access to a Generative AI service. 
The following are supported:

  • OCI Generative AI Service
  • Open AI
  • Cohere

To gain the most from this presentation, attendees should be familiar with basic Oracle Database technology as well as basic AI/ML and LLM concepts.

Other Workshops you might like