Home

websickit is a simple httpserver written in python which can hold multiple machine learning models. The models are registered at webscikit.conf, which is a json-file and looks like this:

{ "/boston": "boston_model.pkl", "/boston2": "boston2_model.pkl" }

This means:

• load the model saved in the file "boston_model.pkl" at url "/boston" at server startup.
• load the model saved in the file "boston2_model.pkl" at url "/boston2" at server startup.

When the server starts, it reads the models into main memory and can serve the predictions at the given url.

Example requests can be found at example/requests. All requests for predictions are sent through POST as a json-file at the given url.