Oreilly – Practical Retrieval Augmented Generation (RAG) 2024-10

Published on: 2024-12-15 18:40:07

Categories: 28

Description

Practical Retrieval Augmented Generation (RAG) course. This course teaches you how to improve existing large language models (LLMs) using information retrieval. This method allows models to access additional information that was not part of their original training data.

What you will learn:

Understand the different types of LLMs and how they fit into the Recovery Augmented Recovery (RAG) system.
Building a RAG application using a vector database and multiple embedders
Testing different generators like GPT-4o, Claude, Command-R and more
Building your own API to implement RAG
Show a chat program based on your RAG work
Using advanced techniques such as GraphRAG

Who is this course suitable for:

Developers, data scientists, and engineers interested in improving the output of their large language models.

Course details

Publisher: Oreilly
Instructor: Sinan Ozdemir
Training level: Beginner to advanced
Training duration: 1 hour and 56 minutes

Course headings

Introduction
1. Practical Retrieval Augmented Generation (RAG): Introduction
Lesson 1: Introduction to Retrieval-Augmented Generation
1. Topics
2. 1.1 Overview of RAG Concepts
3. 1.2 The Family Tree of Large Language Models (LLMs)
4. 1.3 Key Components: Retrievers and Generators
Lesson 2: Building the Foundations
1. Topics
2. 2.1 Introduction to Semantic Search
3. 2.2 Implementing a Simple Indexer/Retriever
Lesson 3: Advanced Prompt Engineering Techniques
1. Topics
2. 3.1 Crafting Effective Prompts
3. 3.2 Few-Shot Learning and Chain-of-Thought Prompting
4. 3.3 Designing RAG Prompts for Consistency
Lesson 4: Developing a RAG System
1. Topics
2. 4.1 Building a RAG Chatbot with GPT-4
3. 4.2 Testing Different LLMs for Retrieval and Generation
Lesson 5: Evaluation and Testing of RAG Systems
1. Topics
2. 5.1 Evaluating the Retriever Part 1
3. 5.2 Evaluating the Retriever Part 2
4. 5.3 Assessing Generative Responses
Lesson 6: Expanding and Applying RAG Systems
1. Topics
2. 6.1 Fine-Tuning Open-Source Embedders with Synthetic Data
3. 6.2 Extending RAG Systems with Re-ranking
4. 6.3 Graph DB + RAG == GraphRAG
5. 6.4 Developing a GraphRAG System
Summary
1. Practical Retrieval Augmented Generation (RAG): Summary

Course prerequisites:

Proficiency in Python 3 with some experience working in interactive Python environments such as Jupyter Notebook, Google Colab, or Kaggle Cores
Introduction to Pandas and Python libraries
Understand basic machine learning/deep learning concepts including training/testing partitioning, loss/cost functions, and gradient descent