Music Inpainting Benchmark

Benchmark of existing inpainting solutions:

Models

Model	Year	Repo	Paper
DeepBach	2017	Repo	Paper
CocoNet	2017	Repo	Paper
AnticipationRNN	2018	Repo	Paper
InpaintNet	2019	Repo	Paper
Music SketchNet	2020	Repo	Paper
Variable Length Infilling	2021	Repo	Paper

Datasets

Dataset	Size	Description	Source	Paper	Type
AILabs	1747	Live Piano Performances	Source	Paper	Single Instrument Polyphony
JSB Chorales	385	Bach Chorales Scores	Source	-	Fixed Voices Polyphony
IrishFolk	45849	Irish Folk Songs	Source	Paper	Monophony

Data Representation

Music SketchNet

DEFAULT_FRACTION: 24

# 0-127 note, 128 hold, 129 rest
note_seq: [
    [48, 128, 128, 128, 128, 128, 50, 128, 128, 128, 128, 128, 52, 128, 128, 128, 128, 128, 53, 128, 128, 128, 128, 128]
]

# [px, rx, len_x, nrx, gd]
factorized: [
    [48, 50, 52, 53, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128, 128],
    [1, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2],
    [4],
    [[1, 0, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [1, 0, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [1, 0, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [1, 0, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0], [0, 1, 0]],
    [48, 128, 128, 128, 128, 128, 50, 128, 128, 128, 128, 128, 52, 128, 128, 128, 128, 128, 53, 128, 128, 128, 128, 128]
]

model_input: [n_batch, **REVISAR**]
model_output: [n_batch, n_measures_middle, DEFAULT_FRACTION, n_classes]

DeepBach

index2note: 
    {0: 'D#5', 1: 'E-4', 2: 'E-5', 3: 'rest', 4: 'F#4', 5: 'E#5', 6: 'G#4', 7: 'B4', 8: 'D4', 9: 'A5', 10: 'END', 11: 'G-4', 12: 'C#5', 13: 'G4', 14: 'A3', 15: 'D#4', 16: 'START', 17: 'D5', 18: 'C5', 19: 'F5', 20: 'A-4', 21: 'C4', 22: 'C#4', 23: 'E5', 24: 'E#4', 25: 'A#4', 26: 'D-5', 27: 'E4', 28: 'G-5', 29: 'A-5', 30: 'A4', 31: 'G5', 32: 'B-4', 33: 'F#5', 34: '__', 35: 'F4', 36: 'OOR', 37: 'G#5', 38: 'B3'} 

score_tensor = tensor([[36, 34, 34, 34, 36, 34, 34, 34, 34, 34, 34, 34,  8, 34, 34, 34, 38, 34,
         34, 34, 34, 34, 14, 34, 36, 34, 34, 34, 36, 34, 34, 34, 34, 34, 14, 34,
         38, 34, 34, 34, 14, 34, 34, 34, 34, 34, 34, 34, 38, 34, 34, 34,  8, 34,
         34, 34, 34, 34, 34, 34, 21, 34, 34, 34, 38, 34, 34, 34, 14, 34, 34, 34,
         34, 34, 34, 34, 36, 34, 34, 34, 34, 34, 34, 34, 38, 34, 34, 34, 38, 34,
         34, 34, 21, 34, 34, 34,  8, 34, 34, 34,  8, 34, 34, 34, 34, 34, 21, 34,
         38, 34, 34, 34, 14, 34, 34, 34, 34, 34, 34, 34, 36, 34, 34, 34, 38, 34,
         34, 34, 34, 34, 34, 34, 21, 34, 34, 34,  8, 34, 34, 34, 34, 34, 34, 34,
         21, 34, 34, 34, 38, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 34, 36, 34,
         34, 34, 34, 34, 34, 34, 38, 34, 34, 34,  8, 34, 34, 34, 34, 34, 34, 34,
         21, 34, 34, 34, 38, 34, 34, 34, 34, 34, 34, 34, 14, 34, 34, 34, 36, 34,
         34, 34, 34, 34, 14, 34, 38, 34, 34, 34, 14, 34, 34, 34, 34, 34, 34, 34,
         38, 34, 34, 34,  8, 34, 34, 34, 34, 34, 34, 34, 21, 34, 34, 34, 38, 34,
         34, 34, 14, 34, 34, 34, 34, 34, 34, 34, 36, 34, 34, 34, 34, 34, 34, 34]])

# Metadata = [Fermata, Tick, Key, N_Voice]
metadata_tensor = tensor([[ 0,  0, 15,  0],
        [ 0,  1, 15,  0],
        [ 0,  2, 15,  0],
        [ 0,  3, 15,  0],
        [ 0,  0, 15,  0],
        [ 0,  1, 15,  0],
        [ 0,  3, 15,  0],
        [ 0,  0, 15,  0],
        [ 1,  2, 15,  0],
        [ 1,  3, 15,  0],
        [ 1,  0, 15,  0],
        [ 1,  1, 15,  0],
        [ 1,  2, 15,  0],
        [ 1,  3, 15,  0]])

---

Project Organization

├── LICENSE
├── Makefile           <- Makefile with commands like `make data` or `make train`
├── README.md          <- The top-level README for developers using this project.
├── data
│   ├── external       <- Data from third party sources.
│   ├── interim        <- Intermediate data that has been transformed.
│   ├── processed      <- The final, canonical data sets for modeling.
│   └── raw            <- The original, immutable data dump.
│
├── docs               <- A default Sphinx project; see sphinx-doc.org for details
│
├── models             <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks          <- Jupyter notebooks. Naming convention is a number (for ordering),
│                         the creator's initials, and a short `-` delimited description, e.g.
│                         `1.0-jqp-initial-data-exploration`.
│
├── references         <- Data dictionaries, manuals, and all other explanatory materials.
│
├── reports            <- Generated analysis as HTML, PDF, LaTeX, etc.
│   └── figures        <- Generated graphics and figures to be used in reporting
│
├── requirements.txt   <- The requirements file for reproducing the analysis environment, e.g.
│                         generated with `pip freeze > requirements.txt`
│
├── setup.py           <- makes project pip installable (pip install -e .) so src can be imported
├── src                <- Source code for use in this project.
│   ├── __init__.py    <- Makes src a Python module
│   │
│   ├── data           <- Scripts to download or generate data
│   │   └── make_dataset.py
│   │
│   ├── features       <- Scripts to turn raw data into features for modeling
│   │   └── build_features.py
│   │
│   ├── models         <- Scripts to train models and then use trained models to make
│   │   │                 predictions
|   |   ├── model_src  <- Original implementation of models
│   │   ├── predict_model.py
│   │   └── train_model.py
│   │
│   └── visualization  <- Scripts to create exploratory and results oriented visualizations
│       └── visualize.py
│
└── tox.ini            <- tox file with settings for running tox; see tox.readthedocs.io

---

Project based on the cookiecutter data science project template.