Core of Oríon

Oríon is an asynchronous distributed framework for black-box function optimization.

Its purpose is to serve as a hyperparameter optimizer for machine learning models and training, as well as a flexible experimentation platform for large scale asynchronous optimization procedures.

It has been designed firstly to disrupt a user’s workflow at minimum, allowing fast and efficient hyperparameter tuning, and secondly to provide secondary APIs for more advanced features, such as dynamically reporting validation scores on training time for automatic early stopping or on-the-fly reconfiguration.

Start by having a look here:


Define the config and fill it based on global configuration files.


Create and define the fields of the configuration object.


Create and define the fields of the database configuration.


Create and define the fields of the evc configuration.


Create and define the fields of generic experiment configuration.


Create and define the field of frontends URI configuration.


Create and define the fields of the storage configuration.


Create and define the fields of the worker configuration.