Build a simple Retrieval-Augmented Generation (RAG) system from scratch using Llama 3.1, Groq API, Sqlite-vec, and FastEmbed in this comprehensive tutorial video. Learn to create a chatbot with custom data without relying on external libraries like LangChain and LlamaIndex. Explore the process of setting up Google Colab, utilizing sqlite-vec for vector storage, adding custom data to the database, creating document embeddings, and understanding how vectors are stored. Discover techniques for similar document search and constructing essential RAG components. Practice interacting with the chatbot using your custom data and gain insights into building efficient AI-powered information retrieval systems.
RAG from Scratch with Llama 3.1 - Building a Custom Data Chatbot
Overview
Syllabus
- Why build RAG from scratch?
- Text tutorial on MLExpert.io
- Google Colab Setup
- sqlite-vec
- Add custom data to the database
- Create document embeddings
- How vectors are stored in the database
- Similar document search
- Build components for our RAG
- Asking the chatbot about our custom data
- Conclusion
Taught by
Venelin Valkov
Related Courses
-
Building and Deploying a RAG Chatbot with JavaScript, LangChain.js, Next.js, Vercel and OpenAI
-
LangChain Chat with Your Data
-
Vector Search RAG Tutorial – Combine Your Data with LLMs with Advanced Search
-
Vector Databases: from Embeddings to Applications
-
RAG From Scratch - Implementing Retrieval-Augmented Generation with LangChain
-
Open-source LLMs: Uncensored & secure AI locally with RAG
