Datasets splits, i.e. closed set and open set, for reproducing our ICASSP21 paper.
We provide the sng_id of the DALI dataset split (https://github.com/gabolsgabs/DALI) used to train models presented in the paper Singing Language Identification Using a Deep Phonotactic Approach (ICASSP 2021).