Model weights for experimental results

These are the model weights trained as reported in the paper.

They are configured with the SE-Resnet34 backbone, followed by linear heads corresponding to the approaches used.

Unzip and paste the folders into src/results/ (refer to the hierarchy example below).

You can use the evaluation command (at README.md/#3-3) to see how it works.

/
├─ src\   
  ├─ results\
    ├─ DCASE2020-ADYOLO\
    ├─ DCASE2021-ADYOLO\
      ...
    └─ DCASE2022-ADYOLO\