Drum Dataset Ecosystem
Collection
Datasets for automatic drum transcription, source separation, and audio synthesis research. • 6 items • Updated • 1
from datasets import load_dataset
ds = load_dataset("schism-audio/waivops-jaz-drm", streaming=True)
Part of the Patchbanks WaivOps collection of AI-generated drum loop datasets.
Each WAV file is paired with a JSON metadata file containing MIDI pitch data, tempo, and time signature labels.
The Hub dataset viewer is configured to load the WAV files as the default split. The paired JSON sidecars remain available in the repository files.
| Split | Audio files |
|---|---|
train |
1,675 |
@dataset{JAZ-DRM Dataset,
author = {WaivOps},
title = {WaivOps JAZ-DRM: Open Audio Resources for Machine Learning in Music},
year = {2025},
doi = {10.5281/zenodo.15658848},
url = {https://doi.org/10.5281/zenodo.15658848},
}
Part of the SchismAudio drum dataset ecosystem.