Huge Symbolic Music Dataset (HSMusic)
HSMusic is a large and participative MIDI collection. More precisely, it is:
- A joint effort to gather as much MIDI files as possible; and automatically tag them.
- A small Python library to feed Machine Learning models with MIDI files.
| Count | 130 943 |
|---|---|
| Data format | MIDI |
| Tags | Style, composer, title and many more (multilabels) |
How to contribute?
If you want to add MIDI files to HSMusic Dataset, please contact me: adrien.pavao@gmail.com. You can also become a contributor of this repository or raise Github issues.
Roadmap
- Complete the data collection
- Clean the code for data management and labeling
- Code ML baselines (discriminative and generative models)
- Write documentation and example notebook
Library overview
- to_matrix: convert MIDI file into a binary matrix
- to_midi: convert a binary matrix into a MIDI file
- Some models...
- TODO: read_data, data_augmentation (transposition, rythm, etc.), change_mode