Planned intervention: On Thursday March 28th 07:00 UTC Zenodo will be unavailable for up to 5 minutes to perform a database upgrade.
Published August 22, 2018 | Version v1
Other Open

Adversarial Unsupervised Domain Adaptation for Acoustic Scene Classification

  • 1. Tampere University of Technology
  • 2. Montreal Institute for Learning Algorithms

Description

The two .zip files contain the pre-trained weights, features, and labels for the AUDASC

In order to use them, you need the code of AUDASC (available from here). The code of AUDASC is based on PyTorch framework

For easy and efficient reproducibility, we include our extracted features and labels (one-hot encoded) from the development dataset of the DCASE 2018 Task 1, subtask B. The license specified by the DCASE 2018 Task1, subtask B, for the data is applied here to the extracted features and labels as well.

Files

AUDASC_features_labels.zip

Files (2.2 GB)

Name Size Download all
md5:8613e7937a923a71a4684cdf543698ad
2.2 GB Preview Download
md5:5841f13589dc17259fbfd7ddff047713
8.7 MB Preview Download

Additional details

Related works

Is supplement to
arXiv:1808.05777 (arXiv)