Oríon is built to be easily extensible. Algorithms and database backends can be implemented in external repositories and installed as plugins. The installation process is very simple, you only need to install them as you would for Oríon’s core (see Installing Oríon). They will be automatically detected and available for Oríon at run-time.

For more information about how to develop your own plugins, see section Developing algorithms.


Note that active development is currently focused on the core. Therefore only one algorithm plugin is available for now.

Skopt algorithms

orion.algo.skopt provides a wrapper for Bayesian optimizer using Gaussian process implemented in scikit optimize. For more information about algorithms configuration and usage please refer to Algorithms.

pip install orion.algo.skopt

Next we define the file bayes.yaml as this

    name: orion-with-bayes
    algorithm: BayesianOptimizer

Then call orion hunt with the configuration file.

$ orion hunt --config bayes.yaml ./ --lr~'loguniform(1e-5, 1.0)'

Now, we have a Bayesian optimizer sampling learning-rate values to optimize the error rate.