In this course, we will explore various developments in RAG systems. The goal is to empower learners to understand, build, and implement RAG systems in various domains (text, multimodal, agentic) with reusable code, blogs, and practical examples.
To download a copy of this repository, execute the following command in your terminal:
git clone --depth 1 https://github.com/marcharaoui/RAG-from-scratch.git| Chapter Title | Overview | Quick Access | Directory |
|---|---|---|---|
| Ch 1: Introduction: Understanding RAG | - Key Limitations of LLMs - Why RAG Matters - How RAG Works |
- Read blog | ./chapter1 |
| Ch 2: Text-only RAG | - Indexing Knowledge: How to organize your data for efficient retrieval - The Art of Chunking: Breaking documents into manageable pieces without losing context - Embedding Models & Advanced Indexing: Choosing the right models and techniques for top-tier performance - Search Strategies: From keyword and semantic search to hybrid and filtered vector search - Augmented Prompt Construction: Crafting smart prompts for generative steps |
- Read blog -Basic RAG notebook |
./chapter2 |
| Ch 3: Multimodal RAG | Coming soon | Coming soon (General, MaxSim, ColPali, full code DIY) | Coming soon |
| Ch 4: Agentic RAG | Coming soon | Coming soon (General, routing, smart query, advanced processing, full code DIY) | Coming soon |
| Ch 5: Bonus | Coming soon | Coming soon (Rerank, evaluation, other techniques) | Coming soon |
This is an ongoing project, the table of contents could possibly change over time.
- Blogs: Theoretical insights, walkthroughs, and narratives. Published on Medium.
- Code Repositories: Modular, well-documented, and interactive codes for hands-on learning, published for exploring and deployment.
- Social Media Outreach: Share summaries, insights, and progress updates on LinkedIn and X.
Enjoy the read 🤗
Note: This entire course is co-authored by both Marc Haraoui and LLM technology.