The demo page includes two models, Composer and Lyricist.
Composer Model (ver. 0.1): Melodies from random seeds.
Lyricist Model (ver. 0.2): 2-3 lines of lyrics from users’ keywords.
The results from A.I. model are references to help musicians in sketch stage of making music. Our A.I. model provides melodies for vocal line that matches musicians’ instrumental music. It creates ready-to-use references from users’ input.
Our model learns from tons of melodies and lyrics from all genres, country, and style. Based on this big data, the model attempts to play variation. Every time It produce different outcome, even from the same input.
Making music is a very time-consuming process. A.I. technology can reduce time and resources of making music, without sacrificing diversity. It would make the process of creation more enjoyable and productive.