/roberta-qa-agent

Simple QA AI agent

Primary LanguagePython

Simple QA chain

This project utilizes paraphrase-MiniLM-L6-v2 and roberta-base-squad2 to power a simple question-answering (QA) system.

Overview

This QA system leverages the capabilities of state-of-the-art NLP models to provide accurate and relevant answers to user queries. It is designed to be easy to set up and use, making it ideal for a variety of applications where quick and reliable information retrieval is key.

Setup Instructions

Environment Setup:

  1. Create a .env file in the root directory of the project. This file should contain the following environment variables:
PINECONE_ENV=<your_pinecone_environment>
PINECONE_API_KEY=<your_pinecone_api_key>
PINECONE_INDEX=<your_pinecone_index_name>
  1. Create a kb.json file. This file should contain the knowledge base that the chatbot will use to answer questions. Ensure it follows the required JSON format.
  2. Run the python setup.py. This will generate and insert vectors into your Pinecone database, which are necessary for the chatbot to understand and retrieve information from the knowledge base.
  3. To ask a question, use the command line interface: python main.py "<your question>"