Skip to content

ajkh88/BassGAN

Repository files navigation

BassGAN

A Generative Adversarial Network (GAN) model for generating bass lines from musical stems.

Setup

pip install -r requirements.txt

Usage

  • See MoisesDB for details on downloading the dataset.
  • Once the dataset is downloaded, set the MOISES_DATA_PATH in prep data to point at the directory you downloaded to.
  • Set the CHUNK_FILES_DIR to the output dir from the prep data job.
  • Use the following command to run the training:
python train.py

See train.py for more details of flags and hyperparameters.

For best results, use the Lightning.ai platform and a GPU.

About

GAN to generate Bass stems from contextual audio

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages