ZnTrack: DVC Machine Learning Pipelines in Python and Jupyter Notebooks

Fabian Zills, PhD Student at the University of Stuttgart, presents on ZNTrack, a way to create, run, and benchmark DVC Pipelines in Python. From the repo: "ZnTrack [zɪŋk træk] is an easy-to-use package for tracking parameters in your Python projects. What is a parameter? Anything set by a user in your code, for example, the number of layers in a neural network layer or the window size of a moving average. ZnTrack works by storing the values of parameters in Python classes and functions and monitoring how they change for several different runs. These changes can then be compared graphically to see what effect they had on your workflow. Beyond the standard tracking of parameters in a project, ZnTrack can be used to deploy jobs with a set of different parameter values, avoid the re-running of components of code where parameters have not changed, and to identify computational bottlenecks in your code." *Additional Links* Fabian Zills Twitter: https://twitter.com/PythonFZ Example Repo: https://github.com/PythonFZ/DVCExample/tree/mnist_sign_language Slides: https://1drv.ms/p/s!AnAGevdAEu1_kmK6AdeMoeHu4vNM?e=o8j42K Binder notebook: https://mybinder.org/v2/gh/PythonFZ/DVCExample.git/mnist_sign_language?labpath=Workflow.ipynb ZnTrack Repo: https://github.com/zincware/ZnTrack Iterative Awesome Projects: https://github.com/iterative/awesome-iterative-projects ------- 📌 START LEARNING DVC • 📘 Docs: https://dvc.org/doc • 🎓 Free Online Course: https://learn.dvc.org • 💻 VS Code Extension: https://marketplace.visualstudio.com/items?itemName=Iterative.dvc • 📺 Subscribe to our channel: https://www.youtube.com/@dvcorg8370?sub_confirmation=1 💬 LET'S CONNECT • GitHub: https://github.com/iterative/dvc • Discord Community: https://dvc.org/chat • Twitter/X: https://twitter.com/DVCorg • LinkedIn: https://www.linkedin.com/company/iterative-ai/ 👇 COMMENT: What part of DVC would you like us to cover next? #DataVersionControl #MLOps #MachineLearning #ReproducibleML #DVC