Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 

Score- and Lyrics-Free Singing Voice Generation

This repository contains the generation code for our paper "Score and Lyrics-Free Singing Voice Generation". The goal is to generate singing voices unconditionally or given an accompaniment track.

You can find the paper at: Score- and Lyrics-Free Singing Voice Generation

The model is implemented with PyTorch.

Note that we have a newer version for unconditional audio generation. You can find the repo at: Unconditional Audio Generation with GAN and Cycle Regularization.

Getting Started

Install requirements

pip install -r requirements.txt

Generate singing voices

Accompanied singing

python generate_singing.py --condition 5 --cuda_id 0

Free singing

python generate_singing.py --condition 0 --cuda_id 0

Vocoder

For converting mel-spectrograms to audios, we use the an WaveRNN version that is implemented by fatchord.

Training scripts

The training scripts will need training data which are not included in this repository.

./scripts/train.accompanied_singer.py
./scripts/train.free_singer.py

About

Score- and Lyrics-Free Singing Voice Generation

Resources

Releases

No releases published

Packages

No packages published

Languages