Website: http://modeldb.csail.mit.edu

ModelDB is an end-to-end system to manage machine learning models. It can ingest models and metadata as they are built, store this data in a structured format, and surface it through a web-frontend for rich querying.

ModelDB frontend

2017.02.08: ModelDB publicly available! Try it out and contribute.

ModelDB currently works with machine learning worklows in spark.ml and scikit-learn.

• Getting Started Guide for spark.ml
• Getting Started Guide for scikit-learn

These guides include instructions on setup, building ModelDB, and incorporating ModelDB into ML workflows. Start there!

We're happy to help with all things ModelDB! Reach out to modeldb_@_lists.csail.mit.edu or post in the ModelDB Google Group.