A smart and responsive Gen AI app that answers your questions based on a given paragraph using Hugging Face Transformers — built with Gradio and perfect for real-time document-based QA!
- Pipeline:
question-answering - Model:
deepset/roberta-base-squad2 - Task: Extractive Question Answering (answers pulled directly from context)
- Framework: Transformers by Hugging Face
- Real-time answer extraction from user-provided context
- Context-aware — only answers if information exists in the paragraph
- Clean, interactive Gradio UI for instant feedback
- Ideal for education, legal/medical docs, support tools & more
This model only answers questions directly related to the input context.
If your question is unrelated or slightly off-topic, the model may give wrong or biased answers.
✅ Example:
- Context: “The Amazon River flows through Brazil, Peru, and Colombia.”
- Question: “Which countries does the Amazon River flow through?”
- → ✅ Answer: Brazil, Peru, and Colombia
❌ Asking general questions like “What is AI?” won’t work unless the context contains the answer.
👉 Try the live demo here: Hugging Face Space
Install dependencies:
pip install -r requirements.txt