Skip to content

Latest commit

 

History

History
28 lines (25 loc) · 1.39 KB

README.md

File metadata and controls

28 lines (25 loc) · 1.39 KB
Raw

Transformer4bats

Introduction

Detection pipeline Recognition pipeline

This repository contains a demonstration of the trained models proposed in "Bat Echolocation Call Detection and Species Recognition by Transformers with Self-Attention".

Setup

  1. Clone this repository and install the required packages. Make sure to use Python 3.7 or newer. 
    cd transformer4bats/ 
pip install -r requirements.txt
  2. Download the model files and place them into the models directory.
  3. Use simple_infer.ipynb to run the bat call detection. By default, a file provided by the Tierstimmenarchiv is used.

Citation

@inproceedings{bellafkir2022bat,
  author={Hicham Bellafkir and Markus Vogelbacher and Jannis Gottwald and Markus M\"uhling and Nikolaus Korfhage and Patrick Lampe and Nicolas Frie\ss{} and Thomas Nauss and Bernd Freisleben},
  title={{Bat Echolocation Call Detection and Species Recognition by Transformers with Self-Attention}},
  year=2022,
  booktitle={The International Conference on Intelligent Systems and Patterns Recognition, Hammamet, Tunisia},
  pages={},
  doi={}
}