Mirdata: Standardized Python Loaders for Music Information Retrieval

Mirdata provides standardized Python loaders for Music Information Retrieval (MIR) datasets, maintained by the mir-dataset-loaders organization with updates through February 2026. It enables programmatic access to audio files and musical annotations such as beats, chords, and melodies across various research collections.

Use Cases

• Training rhythm models using `beats` and `tempo` annotation attributes
• Developing harmonic analysis tools using standardized `chords` labels
• Evaluating pitch tracking algorithms using time-series `f0` or `melody` data

Strengths

• BSD-3-Clause license
• Standardized API for multiple datasets
• Automated data validation using checksums

Limitations

• Does not host raw audio files directly
• Requires Python environment
• Dataset coverage is limited to implemented loaders

Provenance

Source

mir-dataset-loaders GitHub repository

Collection Method

Aggregated loaders for existing research datasets

Freshness

Updated as of February 2026