mirdata

This library provides tools for working with common MIR datasets, including tools for:

• downloading datasets to a common location and format
• validating that the files for a dataset are all present
• loading annotation files to a common format, consistent with the format required by mir_eval
• parsing track level metadata for detailed evaluations.

This libary was presented in our ISMIR 2019 paper

Installation

pip install mirdata

Documentation

• ⭐ Table of supported datasets ⭐
  • Annotation Type Descriptions
• Examples
  • Basic Example
  • Using mirdata with local vs. remote data
  • Using mirdata with tf.data.Dataset
• API documentation
  • Dataset Loaders
  • Utilities
• FAQ
  • How do I add a new loader?
  • How do I get access to a dataset if the download function says it’s not available?
  • Can you send me the data for a dataset which is not available?
  • How do I request a new dataset?
  • What do I do if my data fails validation?
  • How do you choose the data that is used to create the checksums?
  • Does mirdata provide data loaders for pytorch/Tensorflow?
  • Why didn’t you release a version of this library in MATLAB/C/Java/R?
  • A download link is broken for a loader’s .download() function. What do I do?
  • Why the name, mirdata?
  • If I find a mistake in an annotation, should I fix it in the loader?
  • Does mirdata support data which lives off-disk?

Contribute

• Issue Tracker
• Source Code