mlflow 0.2.0

MLflow: An ML Workflow Tool

Note: The current version of MLflow is an alpha release. This means that APIs and data formats are subject to change!

Note 2: We do not currently support running MLflow on Windows. Despite this, we would appreciate any contributions to make MLflow work better on Windows.

Installing

Install MLflow from PyPi via pip install mlflow

MLflow requires conda to be on the PATH for the projects feature.

Documentation

Official documentation for MLflow can be found at https://mlflow.org/docs/latest/index.html.

Running a Sample App With the Tracking API

The programs in example use the MLflow Tracking API. For instance, run:

python example/quickstart/test.py

This program will use MLflow Tracking API, which logs tracking data in ./mlruns. This can then be viewed with the Tracking UI.

Launching the Tracking UI

The MLflow Tracking UI will show runs logged in ./mlruns at http://localhost:5000. Start it with:

mlflow ui

Running a Project from a URI

The mlflow run command lets you run a project packaged with a MLproject file from a local path or a Git URI:

mlflow run example/tutorial -P alpha=0.4

mlflow run [e-mail]:databricks/mlflow-example.git -P alpha=0.4

See example/tutorial for a sample project with an MLproject file.

Saving and Serving Models

To illustrate managing models, the mlflow.sklearn package can log scikit-learn models as MLflow artifacts and then load them again for serving. There is an example training application in example/quickstart/test_sklearn.py that you can run as follows:

$ python example/quickstart/test_sklearn.py
Score: 0.666
Model saved in run <run-id>

$ mlflow sklearn serve -r <run-id> model

$ curl -d '[{"x": 1}, {"x": -1}]' -H 'Content-Type: application/json' -X POST localhost:5000/invocations

Contributing

We happily welcome contributions to MLflow. Please see our contribution guide for details.