NVIDIA Generative AI Examples
A collection of reference workflows, tutorials and sample applications for building generative-AI systems, such as retrieval-augmented generation pipelines and agentic workflows, optimized for NVIDIA's accelerated software stack.
NVIDIA Generative AI Examples is a starting point for integrating generative-AI systems with the NVIDIA software ecosystem, covering retrieval-augmented-generation pipelines, agentic workflows and model fine-tuning. Data Flywheel tutorials show tool-calling fine-tuning, inference, evaluation and guardrailing using NeMo Microservices and NIM microservices; Safer Agentic AI tutorials show auditing language models for vulnerabilities to unsafe prompts and running inference through multiple guardrail layers in parallel.
Other examples include a GPU-accelerated knowledge-graph RAG pipeline built on the RAPIDS ecosystem, an agentic RAG pipeline using Llama 3.1 with NVIDIA NeMo Retriever, and a collection of notebooks, sample code and reference applications built with vision-capable NIM microservices.