AI GitHub
MLflow for tracking experiments and model artifacts
Record parameters, metrics and artifacts alongside model runs.
Key Takeaways
- A tracking run connects logged values with artifacts and the code's experiment context.
- Choose where tracking metadata and artifacts will live.
- Log two runs with one controlled change, then reproduce an artifact from its recorded configuration.
- Metadata and artifact storage can use different systems.
How it works
A tracking run connects logged values with artifacts and the code's experiment context. Model packaging and registry workflows can carry selected results into later stages.
What you need
Choose where tracking metadata and artifacts will live. Check that a clean client can read the selected run and load only the artifacts it should access.
What to try first
Log two runs with one controlled change, then reproduce an artifact from its recorded configuration. Test loading the selected model in an isolated environment.
Try exporting a model from the tracking workflow into its target runtime. Preserve enough dependency and input-schema information to repeat the check later.
Where to look for failures
Metadata and artifact storage can use different systems. Keep credentials, access control and retention consistent across both.
Related Franklin pages
Choose what needs versioning and what needs tracking. A run's curves are easier to reproduce when they link to the exact data, configuration and saved artifact.
- Log experiment metrics and inspect curves and metadata across runs. Aim for comparing training runs
- Connect Python experiment and data workflows to a DagsHub repository. DagsHub Client for connecting ML projects to DagsHub
For the steps that prepare raw data for training or evaluation, see Franklin’s glossary. Data preprocessing
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗