zxjte9411 / FMA-C-DataSet-for-Vocal-Detection

Vocal detection data sets for deep learning.

DataSet-for-Vocal-Detection

The 2s music dataset is related of the FMA: A Dataset For Music Analysis, a dump of the Free Music Archive (FMA).

Data

1. FMA-C-1 DataSet : 12,254 tracks for training, 4,085 tracks for testing.
2. FMA-C-2 DataSet : 13,752 tracks for training, 4,583 tracks for testing.
3. InHouseTestData : 8,291 tracks.
4. Instrumental : 6,372 tracks.
5. Jamendo : 26,715 tracks for training, 5,972 tracks for testing.
6. KTV : 6,534 tracks for training, 1,367 tracks for testing.
7. MIR-1k : 9,900 tracks.
8. RapData : 391 tracks.
9. A-Cappella : 7,922 tracks.
10. MIDI_dataset (model use).

---