Osprey is a tool for practical hyperparameter optimization of machine learning algorithms. It's designed to provide a practical, easy to use way for application scientists to find parameters that maximize the cross-validation score of a model on their dataset.
Osprey is actively being developed by researchers around the world, with primary application areas in computational protein dynamics and drug design, and distributed under the Apache License (v2.0). All development takes place on GitHub.
osprey is a command line tool. It runs using a simple :ref:`config file
<config_file>` which sets up the problem by describing the :ref:`estimator
<estimator>`, :ref:`search space <search_space>`, :ref:`strategy <strategy>`,
:ref:`dataset <dataset_loader>`, :ref:`cross validation <cross_validation>`,
and storage for the :ref:`results <trials>`.
To get started, run
osprey skeleton to create an example config file, and
then boot up one or more parallel instances of
If you happen to run into any issues while using Osprey or would like suggest a new feature, please take a moment to read our "Contributing" section.
.. toctree:: :maxdepth: 2 background installation getting_started contributing config_file batch_submission changelog