Skip to content

Latest commit

 

History

History
 
 

notebooks

README.md

NVIDIA RAG Blueprint - API Interaction and Deployment Notebooks

Overview

This repository contains Jupyter notebooks demonstrating the usage of NVIDIA RAG Blueprint APIs.

Notebooks:

  1. ingestion_api_usage.ipynb: Demonstrates how to interact with the NVIDIA RAG ingestion service, showcasing how to upload and process documents for retrieval-augmented generation (RAG).
  2. retriever_api_usage.ipynb: Illustrates the use of the NVIDIA RAG retriever service, highlighting different querying techniques and retrieval strategies.
  3. launchable.ipynb: A deployment-ready notebook intended for execution within the brev.dev environment.

Setting Up the Environment

To run these notebooks in a Python virtual environment, follow the steps below:

1. Create and Activate a Virtual Environment

python3 -m virtualenv venv
source venv/bin/activate

2. Install Dependencies

Ensure you have JupyterLab and required dependencies installed:

pip3 install jupyterlab

3. Start JupyterLab

Run the following command to start JupyterLab, allowing access from any IP:

jupyter lab --allow-root --ip=0.0.0.0 --NotebookApp.token='' --port=8889 --no-browser

Once running, you can access JupyterLab by navigating to http://<your-server-ip>:8889 in your browser.

Running the Notebooks

  • Open JupyterLab in your browser.
  • Navigate to the desired notebook and run the cells sequentially.

Deployment (Brev.dev)

For deploying launchable.ipynb in brev.dev, follow the platform's instructions for executing Jupyter notebooks within a cloud-based environment selected based on the hardware requirements specified in the launchable.

Notes

  • Ensure API keys and credentials are correctly set up before making API requests.
  • Modify endpoints or request parameters as necessary to align with your specific use case.