How to make your deep learning model available

The idea behind the deep learning module in PAMGuard is to make published deep learning classifiers available to the wider community. Researchers get access to cutting edge classifiers and authors get more exposure, citations and in general make a greater impact after their hard work - win win!

If you want to make your model available there are a few things you will need to do.

  1. Firstly, you will need a deep learning model that works in PAMGuard - see here for more information on how to do that.

  2. Create a settings file (**.psfx file) which shows an example of the deep learning model running with an example sound file which must be available to share.

  3. Create a Markdown document which provides an overview of your model. Use the Shiu/Palmer Right whale model as a template. Note this must include a link to a publication which describes your model in detail.

  4. Archive the model, settings file and example sound file in a public repository such as Zenodo so that it has a DOI.

Once that is all done, send us through everything - we will check it over and, if all OK, publish on the website.