MuseGAN

MuseGAN is a project on music generation. In essence, we aim to generate polyphonic music of multiple tracks (instruments) with harmonic and rhythmic structure, multi-track interdependency and temporal structure. To our knowledge, our work represents the first approach that deal with these issues altogether.

The models are trained with Lakh Pianoroll Dataset (LPD), a new multi-track piano-roll dataset, in an unsupervised approach. The proposed models are able to generate music either from scratch, or by accompanying a track given by user. Specifically, we use the model to generate pop song phrases consisting of bass, drums, guitar, piano and strings tracks. Sample results are available here.

Paper

Hao-Wen Dong*, Wen-Yi Hsiao*, Li-Chia Yang and Yi-Hsuan Yang, "MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment," in AAAI Conference on Artificial Intelligence (AAAI), 2018. [arxiv] [slides]

*These authors contributed equally to this work.

Usage

import tensorflow as tf
from musegan.core import MuseGAN
from musegan.components import NowbarHybrid
from config import *

# initialize a tensorflow session
with tf.Session() as sess:

    ### prerequisites ###
    # step 1. initialize the training configuration
    t_config = TrainingConfig
    
    # step 2. select the desired model
    model = NowbarHybrid(NowBarHybridConfig)
    
    # step 3. initialize the input data object
    input_data = InputDataNowBarHybrid(model)
    
    # step 4. load training data
    path_train = 'train.npy'
    input_data.add_data(path_train, key='train')
    
    # step 5. initialize the museGAN object
    musegan = MuseGAN(sess, t_config, model)

    ### training ###
    musegan.train(input_data)

    ### load and generate samples ###
    # load pretrained model
    musegan.load(musegan.dir_ckpt)
    
    # add testing data
    path_test = 'train.npy'
    input_data.add_data(path_test, key='test')
    
    # generate samples
    musegan.gen_test(input_data, is_eval=True)

Training Data

• tra_phr.npy (7.54 G): A numpy array that contains the training data for phrases. Its shape is (50266, 384, 84, 5).

• tra_bar.npy (4.79 G): A numpy array that contains the training data for bars. Its shape is (127734, 96, 84, 5).