Open Daily: 10am - 10pm | Alley door pickup 10am - 7pm 3038 Hennepin Ave Minneapolis, MN
612-822-4611

  • Open Daily: 10am - 10pm
    Alley-side Pickup: 10am - 7pm

    3038 Hennepin Ave Minneapolis, MN
    612-822-4611

Open Daily: 10am - 10pm | Alley-side Pickup: 10am - 7pm
3038 Hennepin Ave Minneapolis, MN
612-822-4611
ModelDB for Machine Learning Workflows: Practical Experiment Tracking and Reproducibility

ModelDB for Machine Learning Workflows: Practical Experiment Tracking and Reproducibility

Johnson, Robert U.

Paperback

General Computers

Currently unavailable to order

ISBN13: 9798170897223
Publisher: Independently Published
Pages: 206
Weight: 0.62
Height: 0.44 Width: 6.00 Depth: 9.00
Language: English
ModelDB for Machine Learning Workflows: Practical Experiment Tracking and Reproducibility is a comprehensive guide to managing machine learning experiments in a way that is reproducible, collaborative, and scalable. It explains why experiment tracking matters, showing how teams can systematically capture the details behind every run, compare results with confidence, and preserve the context needed to revisit and extend prior work. The book provides a clear foundation in the core ideas, challenges, and design goals of modern experiment management systems.

At its core, the book explores ModelDB's architecture and the metadata model that supports it, including storage design, API structure, graph-based lineage tracking, and security considerations. It walks readers through the practical capture of experiment information such as datasets, code versions, hyperparameters, artifacts, and evaluation metrics, while also discussing the trade-offs involved in building reliable tracking infrastructure. Integration examples demonstrate how ModelDB fits into common machine learning environments, including scikit-learn, TensorFlow, and PyTorch, as well as data versioning tools, CI/CD pipelines, and broader MLOps workflows.

Beyond the fundamentals, the book covers analysis, visualization, scalability, and extensibility, offering guidance on using ModelDB effectively in real-world settings. It also presents case studies, best practices, and lessons learned from production and research deployments, highlighting how systematic experiment tracking can improve productivity and trust in machine learning work. Whether for researchers, engineers, or platform architects, this book equips readers with practical tools for building reproducible ML workflows and supporting long-term innovation.

Also from

Johnson, Robert U.

View All

Also in

General Computers

View All