• About the Notebooks
• Prerequisites
• Running the Notebooks

The notebooks show how to use the langchain-nvidia-ai-endpoints and llama-index-embeddings-nvidia Python packages. These packages provide the basics for developing a RAG application and performing inference either from NVIDIA API Catalog endpoints or a local deployment of NVIDIA microservices.

• You have Python 3 installed.
• Complete the common prerequisites.

1. Export your NVIDIA API key as an environment variable:

   export NVIDIA_API_KEY="nvapi-<...>"
   

2. Create a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate

3. Install JupyterLab in the virtual environment:

   pip3 install jupyterlab

4. Start the JupyterLab server:

   jupyter lab --allow-root --ip=0.0.0.0 --NotebookApp.token='' --port=8889

5. Open a web browser and access http://localhost:8889/lab.

   Browse to the RAG/notebooks directory to open an execute the cells of the notebooks.