This repository belongs to our paper CoarseSoundNet: Building a reliable model for ecological soundscape analysis (Gebhard et al., 2026). The paper is currently under review and will be linked upon acceptance. This repo contains the code we used for training and evaluating our model, with the corresponding configurations and scripts.
For now please be referred to our huggingface page (https://huggingface.co/HearTheSpecies/CoarseSoundNet), in order to use the model, until we are finished with cleaning up this repository and setting up instructions.
WARNING: This repository is still under construction. We are now in the process of adding content, cleaning up, and adding instructions.