Skip to content
A score-informed variant of Wave-U-Net for source separation of choral music
Python
Branch: master
Clone or download
Pull request Compare This branch is 8 commits ahead of f90:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Input
Models
audio_examples MAJOR REVAMP Aug 26, 2018
checkpoints
data
scripts Add script to generate model graph. Aug 14, 2019
.gitignore
CCMixter.xml Initial commit May 9, 2018
Config.py
Datasets.py Read score-informed dataset of chorales. Aug 14, 2019
Evaluate.py Support score-informed training. Aug 14, 2019
LICENSE Create LICENSE Jun 11, 2018
Plot.py Moved plotting code into extra Plot.py function, making training and … Dec 2, 2018
Predict.py NEW: ADDED 44KHz version of best vocal separator (M5 model) that beat… Nov 16, 2018
README.md
Test.py
Training.py
Utils.py
musb_005_angela thomas wade_audio_model_without_context_cut_28234samples_61002samples_93770samples_126538.wav
mypy.ini
requirements.txt MAJOR REVAMP of dataset input - changed to TFRecord format for faster… Jan 24, 2019
waveunet.png Added requirements May 28, 2018

README.md

Score-informed Wave-U-Net

This repository extends Wave-U-Net to support score-informed separation. It was created as part of my Master's thesis, in which I have used it for source separation of choral music. See the original Wave-U-Net repository for more information on the model architecture and installation instructions.

You can’t perform that action at this time.