The dataset viewer is not available for this split.
Rows from parquet row groups are too big to be read: 2.22 GiB (max=286.10 MiB)
Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

A modified version of the Auditory Attention Detection Dataset KULeuven. Original data can be downloaded here.

This dataset is used for Neuro-steered speaker extraction conditioned on EEG signals in the NeuroHeed paper, which the code can be found here (ClearVoice repo) or here (Paper repo).

Downloads last month
58
Edit dataset card