# midi_melody_extraction

**Repository Path**: ByteDance/midi_melody_extraction

## Basic Information

- **Project Name**: midi_melody_extraction
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2022-12-07
- **Last Updated**: 2026-03-12

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

This repository contains a basic version of our code related to the model presented in the ISMIR2022 paper:  
[A deep learning method for melody extraction from a polyphonic symbolic music representation](https://archives.ismir.net/ismir2022/paper/000091.pdf).  
The data corpus that it is based on can be found [here](https://github.com/music-x-lab/POP909-Dataset/tree/master/POP909).   

## Main end-to-end script
Run `bash run.sh` 

## How to process the features
Run process_data.py

## How to prepare the data for training
Run create_stats_dict.py

## How to Train a melody extraction model
Run train.py

## How to run inference
Run predict.py