Usage
=====

Quick start
-----------

For a full walkthrough, see `pipeline <pipeline.html>`_ and `data generators <data_generators.html>`_.

Typical workflow
----------------

1. Prepare the required data files.
2. Generate shell datasets.
3. Train or load a model from ``experiments/cnnShell_HO/`` or ``experiments/cnnShell_CV/``.
4. Run inference or detection-map scripts.
5. Inspect the results and plots.