Description:
A document-based QnA system that allows users to upload text documents (PDF, TXT, DOCX) and ask natural language questions. The system understands document context using LLMs and delivers precise answers.
Key Features:
Frontend with user-friendly document upload and input UI
Uses BERT for embedding-based context matching
GPT2 and BART for generative-style answering
Answers questions even when phrased differently than in the text
Built-in API endpoints for modular integration
