MLFLOW-EDI

Open-source ML lifecycle platform.

Definition

MLflow v2.0 (November 2022) adds MLflow Pipelines, MLflow LLM (GenAI). MLflow Recipes for templated MLOps. Native integrations PyTorch, scikit-learn, TensorFlow, XGBoost, Spark MLlib, OpenAI. Backend store: SQLite, PostgreSQL, MySQL ; artifact store: S3, Azure Blob, GCS.

Origin

Created by Matei Zaharia and team at Databricks ; open-sourced June 2018 ; v1.0 in June 2019.

Example in context

An ML team logs 200 XGBoost experiments to predict EDI partner delays via MLflow Tracking.

Related terms

• Kubeflow — K8s competitor.