Questgen.ai is an open-source NLP library designed for easy-to-use question generation from text. It targets developers and researchers looking to automate the creation of educational materials, quizzes, or data augmentation sets, leveraging state-of-the-art transformer models.

How It Works

Questgen.ai utilizes multiple T5 transformer models, each fine-tuned for specific question generation tasks: boolean (Yes/No), multiple-choice questions (MCQs), general FAQs, paraphrasing, and question answering. For MCQs, it employs the sense2vec library to generate plausible distractors (incorrect options) based on word embeddings, enhancing the quality of generated questions.

Quick Start & Requirements

• Install:

  pip install git+https://github.com/ramsrigouthamg/Questgen.ai
  pip install git+https://github.com/boudinfl/pke.git
  python -m nltk.downloader universal_tagset
  python -m spacy download en
  

• Dependencies: nltk, spacy, pke, sense2vec (requires downloading word vectors: s2v_reddit_2015_md.tar.gz).
• Demo: https://questgen.ai/

Highlighted Details

• Supports generation of Multiple Choice Questions (MCQs), Boolean Questions, General FAQs, and question paraphrasing.
• Includes functionality for both simple and boolean question answering.
• Leverages T5 transformer models for question generation and answer prediction.
• Uses sense2vec for generating multiple-choice options.

Maintenance & Community

• The project is primarily maintained by Ramsrigoutham Govindarajulu.
• Community links (Discord/Slack, roadmap) are not explicitly mentioned in the README.

Licensing & Compatibility

• The README does not specify a license. This requires further investigation for commercial use or closed-source integration.

Limitations & Caveats

• The project's license is not specified, which may pose a barrier to commercial adoption.
• The README does not detail specific model sizes or hardware requirements for running the T5 models, though transformer models typically require significant computational resources.