CNN usage

SOAP is usually run as a pipeline on the LIGO detectors, both searching for astrophysical signals and as a detector characterisation tool.

The machine learning part of this method, mostly using CNNs to reduce the impact of instrumental artefacts, and how to use it is described below.

These tools were designed to run on the LIGO clusters with access to LIGO data, however may still work elsewhere.

Training a model

Once the data has been generated, its time to train a model on this data. There are a number of ways to do this, however, we will use the same configuration file as on the data_generation page.