Implementation of Pyfilesystem for BirdNET: run BirdNET on HPC for faster results

This repository provide the means of using BirdNET on HPCs (i.e. supercomputers) by leveraging Pyfilesystem. If you use BirdNET please cite:

@article{kahl2021birdnet,
  title={BirdNET: A deep learning solution for avian diversity monitoring},
  author={Kahl, Stefan and Wood, Connor M and Eibl, Maximilian and Klinck, Holger},
  journal={Ecological Informatics},
  volume={61},
  pages={101236},
  year={2021},
  publisher={Elsevier}
}

How to run the scripts

On HPC:

Pull the image as a .sif file for compliance with singularity

singularity pull birdnet.sif ghcr.io/ninanor/birdnet-analyzer:main

Provided the config file has been updated and that the HPC uses TORQUE run:

qsub run_birdnet_hpc.sh

Locally

Pull the docker image:

docker pull ghcr.io/ninanor/birdnet-analyzer:main

Provided the config has been updated (especially CONNECTION_STRING set to False) run:

./run_birdnet_local.sh