MLflow

Use it when

• track model experiments, artifacts, metrics, and model versions
• compare training runs without spreadsheets and zip files

What it solves

Not the fit when

• dataset lineage requires explicit logging or another versioning layer

Install

pip install mlflow

Invoke

Start `mlflow server`, wrap training with an MLflow run, log parameters metrics and artifacts, then register or compare selected models.

Alternatives

No reviewed alternatives recorded yet.