Skip to content

NilsGrossepieper/gradeyourexam

Repository files navigation

gradeyourexam.com

gif_gradeyourexam
Demonstration of gradeyourexam.com


gradeyourexam.com was a university course project that enabled teachers to create, upload, and grade exams automatically using AI-powered evaluation methods.
The project required hosting two large model (Llama 3 8B and all-MiniLM-L6-v2), which made long-term deployment costly.
As a result, the live site is no longer running, but the repository and documentation remain for reference.


📌 What this is

  • Web-based platform for automated exam grading
  • AI-powered similarity scoring between answers and sample solutions
  • Flexible grading schemes including LLM-based evaluation
  • University course project (archived)

🎯 What it does

  • Allows teachers to create and upload exams
  • Compares student answers to sample solutions using embeddings
  • Supports LLM-based and similarity-score-based grading methods
  • Generates downloadable score sheets

🚀 How to use

Requirements: Python 3.10+

Install dependencies:

pip install -r requirements.txt

Run the application:

python main.py

💡 The application will start a local server — open the displayed URL in your browser to access the platform. Note: Hosting both a Llama 3 8B and a all-MiniLM-L6-v2 model requires significant resources; you may need to adjust the configuration for local testing.


🌐 Project Link

Originally accessible at gradeyourexam.com (now offline).
Developed as part of the university course Data Science Project.


👥 Author / Contributors

  • Aizhan Suleimen
  • Nils Großepieper

📜 License

MIT License

About

AI-powered exam grading platform for teachers, with multiple NLP-based scoring methods including LLMs, semantic similarity, and regression models.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Contributors