Platform for Complete Machine Learning Lifecycle
summary
Open source MLflow from Databricks claims an end-to-end ML platform by providing concrete APIs and formats: a tracking API that logs params, metrics and artifacts, a project packaging (Conda/docker entrypoints), and model flavors for TensorFlow/Keras, scikit‑learn and Spark. A short demo trains three Keras IMDB experiments (TensorFlow backend), compares runs in the MLflow UI, picks a model (~90% accuracy), packages it with its Conda environment and reproduces and deploys it. Version 0.8 is available.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 2:47ML lifecycle complexity
- 8:07ML dev/deployment challenges
- 10:53MLflow design principles
- 13:37MLflow core components
- 17:30Tracking experiments
- 18:54Packaging projects
- 21:21Deploying models
- 23:07MLflow tracking demo
- 32:34MLflow 0.8 release
- 35:25Open-source community
related talks
31mMLflow: Platform for Complete Machine Learning Lifecycle
24mMLflow: An open platform to simplify...
30mMLflow: an Open Source Tool for Managing the ML Project Lifecycle
57mManaging Machine Learning Lifecycle with MLflow
33mOperationalizing Model Lifecycle with MLFlow
45mMachine Learning Lifecycle Made Easy with MLflow
17mIntroducing mlflow
30mEnd-to-End MLOps with MLflow and Kubeflow