ds2-lab/NotebookOS: NotebookOS - AWS - ASPLOS'26

NotebookOS is a fully-featured prototype designed for GPU-based machine learning training workloads. This platform provides seamless integration with powerful monitoring and analytics tools, making it ideal for data scientists and machine learning engineers. It is also designed to provide a high degree of interactivity, making it well-suited for interactive programming sessions and thus a great tool for students.

Key Features

• GPU-Accelerated Notebooks 🚀: Leverage GPU power for machine learning and deep learning workloads.
• Fully-Featured Jupyter Environment 💻: Provides a robust Jupyter environment for interactive data analysis and model development.
• Integrated Monitoring & Analytics 📈:
  • Prometheus for time-series data monitoring 🔍
  • Promtail for log collection 📝
  • Loki for efficient log aggregation ♻️
  • Grafana for real-time data visualization and log viewing 📊

This artifact contains a copy of the main branch of the source code from the NotebookOS GitHub repository and the NotebookOS Administrative Dashboard & Workload Orchestrator GitHub repository.