NVIDIA/GenerativeAIExamples

Provides end-to-end templates for RAG pipelines, agentic workflows, and data flywheel systems using NVIDIA NIM microservices, NeMo fine-tuning/evaluation tools, and guardrails—integrated with LangChain, LlamaIndex, and Haystack. Covers specialized patterns like knowledge graph RAG with RAPIDS GPU acceleration, vision model pipelines, and tool-calling workflows with safety auditing via NeMo Auditor. Docker Compose examples enable rapid local deployment on NVIDIA infrastructure with cloud API fallback options.

3,839 stars. Actively maintained with 6 commits in the last 30 days.