This project implements a lightweight Retrieval-Augmented Generation (RAG) chatbot focused on answering queries related to US Basic Laws. The chatbot leverages the HuggingFaceTB/SmolLM2-1.7B-Instruct model, chosen for its efficiency and effectiveness in resource-constrained environments.
The system is designed to provide accurate and contextual responses by combining retrieval from a custom knowledge base with generative capabilities. It can be executed seamlessly in a Google Colab environment, making it accessible for experimentation and deployment.
Features Lightweight Model: Efficient inference using SmolLM2-1.7B-Instruct. Knowledge Base Integration: Tailored retrieval mechanism for US Basic Laws. Colab Ready: Easy setup and execution in a Colab notebook. Usage Load the pre-trained model from Hugging Face. Set up the retrieval system with your custom legal document dataset. Start asking questions to receive informed, contextual responses.