×
Suggest this Article to:
TOP INDEXERS
Public Article
-
verified
GENERAL ADVERSARIL NETWORK (GAN) FOR MUSIC GENERATION
ISSN: 2582 - 9130
Publisher: author
GENERAL ADVERSARIL NETWORK (GAN) FOR MUSIC GENERATION
Indexed in
Technology and Engineering
ARTICLE-FACTOR
1.3
Article Basics Score: 2
Article Transparency Score: 2
Article Operation Score: 3
Article Articles Score: 3
Article Accessibility Score: 2
SUBMIT PAPER ASK QUESTION
International Category Code (ICC):
ICC-1802
Publisher: Krishma Publication
International Journal Address (IAA):
IAA.ZONE/2582383999130
eISSN
:
2582 - 9130
VALID
ISSN Validator
Abstract
Most existing brain network models for music age utilize repetitive brain organizations. Be that as it may, the new WaveNet model proposed by DeepMind shows that convolutional brain organizations (CNNs) can likewise create sensible melodic waveforms in the sound area. Following this light, we research involving CNNs for creating tune (a series of MIDI notes) an endless flow of bars in the emblematic area. Notwithstanding the generator, we utilize a discriminator to become familiar with the disseminations of tunes, making it a generative ill-disposed network (GAN). In addition, we propose an original contingent system to take advantage of accessible earlier information, with the goal that the model can create tunes by the same token without any preparation, by following a harmony succession, or by molding on the tune of past bars (for example a preparing tune), among different potential outcomes. The subsequent model, named MidiNet, ca...
