Skip to content

Leverage Retrieval-Augmented Generation (RAG) to analyze new government policies and assess their impact on the organization.

Notifications You must be signed in to change notification settings

vanessabuilds/LeanAI-Policy-Reader

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 

Repository files navigation

LeanAI Policy Reader

A lightweight, cost-effective AI tool for analyzing and summarizing policy documents using IBM's Granite 3.3-8B model.

Overview

The LeanAI Policy Reader is designed to efficiently process and interpret policy documents, providing concise summaries and analyses. By leveraging IBM's Granite 3.3-8B model, the tool ensures high-quality outputs while maintaining resource efficiency.

Key Features

  • Policy Document Analysis: Processes policy URL to extract structured summaries and insights.
  • Granite 3.3-8B Integration: Utilizes IBM's advanced language model for accurate and context-aware interpretations.
  • Cost-Efficient Processing: Optimized for minimal computational resources without compromising performance.
  • User-Friendly Interface: Interactive Jupyter Notebook for seamless user interaction and customization.

IBM Granite 3.3-8B Model

IBM's Granite 3.3-8B is an open-source, 8-billion parameter language model designed for enhanced reasoning and instruction-following capabilities. Key attributes include:

  • Extended Context Length: Supports up to 128K tokens, enabling processing of long documents and conversations.
  • Multilingual Support: Trained on data across 12 languages, including English, German, Spanish, French, Japanese, and more.
  • Advanced Reasoning: Incorporates structured reasoning techniques, such as chain-of-thought prompting, for improved performance on complex tasks.
  • Fill-in-the-Middle (FIM): Supports FIM for code completion tasks, enhancing its utility in programming-related applications.
  • Open-Source License: Released under the Apache 2.0 license, promoting transparency and community collaboration.

For more details, refer to the IBM Granite Documentation.

Acknowledgments

  • IBM Research: For developing the Granite 3.3-8B model and contributing to the open-source AI community.
  • Replicate: For providing an accessible platform to deploy and experiment with machine learning models.

Demo


About

Leverage Retrieval-Augmented Generation (RAG) to analyze new government policies and assess their impact on the organization.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published