mir_eval

Note: This web service is running an out-of-date version of mir_eval and should be treated only as a proof-of-concept.

Use the form below to evaluate annotations for a given MIR task.
The file format should be as described in mir_eval's documentation.
Some example annotation files can be found in within mir_eval's tests.
You can also query this web service as an API, e.g.:
curl -F "task=beat" -F "estimated_file=@est.txt" -F "reference_file=@ref.txt" http://labrosa.ee.columbia.edu/mir_eval/
task should be one of beat, chord, melody, onset, pattern, segment, tempo, key, or transcription.
If you're running a large-scale evaluation, it will probably be more efficient to run mir_eval locally.
Installation instructions for mir_eval can be found here.
You can even run mir_eval with minimal Python knowledge by using the evaluators.
If you use mir_eval in a research project, please cite the following paper:

Colin Raffel, Brian McFee, Eric J. Humphrey, Justin Salamon, Oriol Nieto, Dawen Liang, and Daniel P. W. Ellis.
"mir_eval: A Transparent Implementation of Common MIR Metrics"
Proceedings of the 15th International Conference on Music Information Retrieval, 2014.

Reference file

Estimated file

Task