MLFlow is an open-source Python framework for experiment tracking. The MLFlow plugin for Hamilton includes two sets of features:
- Standard approach to save and load machine learning models
- Automatically track Hamilton pipeline results in MLFlow
This pairs nicely with the Hamilton UI which gives you a way to explore your pipeline code, attributes of the artifacts produced, and execution observability.
👨‍💻 MLFlow plugin tutorial on GitHub: github.com/DAGWorks-Inc/hamil...
🖥 Hamilton UI: hamilton.dagworks.io/en/lates...
0:00 - Introduction
1:00 - Training: train a machine learning model
5:00 - View our trained model in the MLFlow UI
6:18 - Inference: load a machine learning model for predictions
9:00 - Automatically track Hamilton runs in MLFlow
13:30 - View tracked runs in the MLFlow UI
14:13 - Grouping MLFlow runs by Hamilton code version
16:48 - Automated MLFlow model metadata
18:58 - Using Hamilton @tag to add metadata
20:40 - Using MLFlow model aliases
21:22 - Proof that debugging is easy!
22:16 - Tracking runs with Hamilton UI + MLFlow UI
26:40 - Viewing tracked runs in the Hamilton UI
~~~~
⭐Hamilton on GitHub: github.com/DAGWorks-Inc/hamilton
📖 Hamilton documentation: hamilton.dagworks.io/
📣 Join our Slack communtiy: join.slack.com/t/hamilton-ope...