{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/musegan-multi-track-sequential-generative","title":"MuseGAN: Multi-track Sequential Generative Adversarial Networks for Symbolic Music Generation and Accompaniment","arxiv_id":"1709.06298","date":"2017-09-19","proceeding":null,"authors":["Hao-Wen Dong","Wen-Yi Hsiao","Li-Chia Yang","Yi-Hsuan Yang"],"abstract":"Generating music has a few notable differences from generating images and\nvideos. First, music is an art of time, necessitating a temporal model. Second,\nmusic is usually composed of multiple instruments/tracks with their own\ntemporal dynamics, but collectively they unfold over time interdependently.\nLastly, musical notes are often grouped into chords, arpeggios or melodies in\npolyphonic music, and thereby introducing a chronological ordering of notes is\nnot naturally suitable. In this paper, we propose three models for symbolic\nmulti-track music generation under the framework of generative adversarial\nnetworks (GANs). The three models, which differ in the underlying assumptions\nand accordingly the network architectures, are referred to as the jamming\nmodel, the composer model and the hybrid model. We trained the proposed models\non a dataset of over one hundred thousand bars of rock music and applied them\nto generate piano-rolls of five tracks: bass, drums, guitar, piano and strings.\nA few intra-track and inter-track objective metrics are also proposed to\nevaluate the generative results, in addition to a subjective user study. We\nshow that our models can generate coherent music of four bars right from\nscratch (i.e. without human inputs). We also extend our models to human-AI\ncooperative music generation: given a specific track composed by human, we can\ngenerate four additional tracks to accompany it. All code, the dataset and the\nrendered audio samples are available at https://salu133445.github.io/musegan/ .","url_abs":"http://arxiv.org/abs/1709.06298v2","url_pdf":"http://arxiv.org/pdf/1709.06298v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/salu133445/musegan","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/dariaip/musegan_pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/elsalmi/qiskit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/inzva/music-generation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/liuhaumin/leadsheetgan","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/lucylow/Stochastic_SoundCloud","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/safakkbilici/Synthetic-Music-Generation-with-Deep-Neural-Networks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"musegan-multi-track-sequential-generative","repo_url":"https://github.com/ucsd-dsc-arts/dsc160-final-dsc160-final-group19","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"music-generation","task_name":"Music Generation"}],"methods":[],"datasets_introduced":[{"slug":"lakh-pianoroll-dataset","name":"Lakh Pianoroll Dataset","full_name":""}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1709.06298","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}