A Streamlit-based NLP application that intelligently summarizes PDF documents and allows users to ask questions about the content using free Hugging Face transformer models.
- 📄 PDF Upload: Upload and extract text from any PDF using PyMuPDF (fitz).
- 🧾 Text Preview: View extracted text in a scrollable box before processing.
- ✂️ Chunking: Automatically splits long documents into 1000-word chunks for efficient summarization.
- 🧠 Summarization: Generates concise document summaries using
facebook/bart-large-cnn. - ❓ Q&A System: Ask questions about the document — the system retrieves the most relevant text using semantic similarity (
sentence-transformers/all-MiniLM-L6-v2) and answers withdeepset/roberta-base-squad2. - ⚡ Real-time Feedback: Displays progress spinners, success messages, and organized output sections for a smooth user experience.
- 🎨 Clean UI: Built with Streamlit — minimal, modern, and responsive.
- Frontend: Streamlit
- Backend / NLP: Hugging Face Transformers
- Models Used:
- Summarization →
facebook/bart-large-cnn - Sentence Embeddings →
sentence-transformers/all-MiniLM-L6-v2 - Question Answering →
deepset/roberta-base-squad2
- Summarization →
- Text Extraction: PyMuPDF (fitz)
- Language: Python