An Introduction to Retrieval Augmented Generation
summary
Claims RAG lets organisations build chatbots without training LLMs, using retrieval not model fine‑tuning, and the workshop shows that end‑to‑end with code. Tools used are text-embedding-3-small for 1,536‑dim embeddings, a vector store (pinecone) with cosine similarity, LangChain recursive text splitters that chunk ~1,500 characters with 500 overlap, and OpenAI chat completions. The demo loads chunks from the Flat Earth text into a vector index, retrieves top results and generates answers — a working simple RAG chatbot.
AI-written summary · about
chapters (6)
- 0:00Workshop introduction
- 5:09What is RAG
- 17:45RAG tools: vectors & embeddings
- 33:02Chunking text for vectors
- 39:48RAG bot with OpenAI API
- 45:08Exercise and wrap-up
related talks
15mRAG Beyond Chatbots
50mRetrieval Augmented Generation Beyond Toy Demos
51mDeveloping a Custom RAG System
38mRetrieval-Augmented Generation (RAG) in NLP
21mHow RAG is Redefining AI Capabilities
28mWhat the RAG? Demystifying Retrieval Augemented Generation
20mRetrieval Augmented Generation for AI Apps
01h 14mAn Introduction to Retrieval Augmented Generation