Lean RAG Accelerator - Training & QuickStart

This repository contains both training content and deployable quickstart code for the Lean RAG Accelerator.

Repository Structure

lean-rag/
├── README.md                    # This file
├── lean-rag-accelerator/       # Deployable QuickStart code
│   ├── README.md               # QuickStart main documentation
│   ├── bootstrap.sh            # Deployment script
│   ├── Makefile                # Make targets for deployment
│   ├── examples/               # Deployment YAML files
│   ├── docs/                   # QuickStart documentation
│   ├── scripts/                # Utility scripts
│   └── charts/                 # Helm charts
├── modules/                     # Training modules (Antora)
│   ├── module-1-*/            # Module 1: Inference Economics
│   ├── module-2-*/            # Module 2: Model Optimization
│   ├── module-3-*/            # Module 3: High-Performance Serving
│   ├── module-4-*/            # Module 4: Standardized RAG
│   └── module-5-*/            # Module 5: Validation & Benchmarking
└── antora.yml                  # Antora configuration

Two Ways to Use This Repository

1. Training Content (Antora Documentation)

The training modules provide conceptual learning and best practices:

• View Online: The rendered training content is available via GitHub Pages (when configured)
• Build Locally: Use Antora to build and view the documentation locally
• Modules: Cover inference economics, model optimization, serving, RAG implementation, and benchmarking

To view training content:

• Navigate to the rendered documentation (GitHub Pages)
• Or build locally: antora antora-playbook.yml

2. Deployable QuickStart Code

The lean-rag-accelerator/ directory contains ready-to-deploy code:

• README.md: Complete quickstart guide with deployment instructions
• bootstrap.sh: Interactive deployment script
• Makefile: Automated deployment targets
• examples/: Kubernetes YAML files for model optimization, inference serving, and RAG application
• docs/: Business value, architecture, and benchmarking documentation

To deploy the quickstart:

cd lean-rag-accelerator
./bootstrap.sh

Or use the Makefile:

cd lean-rag-accelerator
make help
make deploy-all NAMESPACE=lean-rag-accelerator

Quick Links

Training Content

• Course Home - Start here for training modules
• Module 1: Inference Economics
• Module 2: Model Optimization
• Module 3: High-Performance Serving
• Module 4: Standardized RAG
• Module 5: Validation & Benchmarking

QuickStart Code

• QuickStart README - Main quickstart documentation
• Business Value - ROI and problem statement
• Architecture - Technical architecture
• Deployment Guide - Step-by-step deployment
• Benchmarking - Performance validation

Getting Started

For Training (Learning)

1. Review the training modules in order (Module 1 → Module 5)
2. Follow along with the concepts and best practices
3. Reference the quickstart code for hands-on practice

For Deployment (QuickStart)

1. Navigate to lean-rag-accelerator/ directory
2. Read the QuickStart README
3. Follow the deployment instructions
4. Use bootstrap.sh or Makefile for automated deployment

Prerequisites

• For Training: No prerequisites - the modules explain concepts
• For QuickStart:
  • Red Hat OpenShift AI 3.0
  • Kubernetes cluster with GPU nodes
  • kubectl/oc CLI configured
  • See QuickStart Requirements for details

Support

• Training Questions: Review the training modules
• QuickStart Issues: See Troubleshooting Guide
• General: Refer to the main repository or contact Red Hat support

Related Resources

• Red Hat OpenShift AI 3.0 Documentation
• Reference RAG QuickStart
• KServe Documentation
• vLLM Documentation