YuE: Scaling Open Foundation Models for Long-Form Music Generation

We tackle the task of long-form music generation--particularly the challenging \textbf{lyrics-to-song} problem--by introducing YuE, a family of open foundation models based on the LLaMA2 architecture. Specifically, YuE scales to trillions of tokens and generates up to five minutes of music while maintaining lyrical alignment, coherent musical structure, and engaging vocal melodies with appropriate accompaniment. It achieves this through (1) track-decoupled next-token prediction to overcome dense mixture signals, (2) structural progressive conditioning for long-context lyrical alignment, and (3) a multitask, multiphase pre-training recipe to converge and generalize. In addition, we redesign the in-context learning technique for music generation, enabling versatile style transfer (e.g., converting Japanese city pop into an English rap while preserving the original accompaniment) and bidirectional generation. Through extensive evaluation, we demonstrate that YuE matches or even surpasses some of the proprietary systems in musicality and vocal agility. In addition, fine-tuning YuE enables additional controls and enhanced support for tail languages. Furthermore, beyond generation, we show that YuE's learned representations can perform well on music understanding tasks, where the results of YuE match or exceed state-of-the-art methods on the MARBLE benchmark. Keywords: lyrics2song, song generation, long-form, foundation model, music generation

@article{yuan2025_2503.08638,
  title={ YuE: Scaling Open Foundation Models for Long-Form Music Generation },
  author={ Ruibin Yuan and Hanfeng Lin and Shuyue Guo and Ge Zhang and Jiahao Pan and Yongyi Zang and Haohe Liu and Yiming Liang and Wenye Ma and Xingjian Du and Xinrun Du and Zhen Ye and Tianyu Zheng and Yinghao Ma and Minghao Liu and Zeyue Tian and Ziya Zhou and Liumeng Xue and Xingwei Qu and Yizhi Li and Shangda Wu and Tianhao Shen and Ziyang Ma and Jun Zhan and Chunhui Wang and Yatian Wang and Xiaowei Chi and Xinyue Zhang and Zhenzhu Yang and Xiangzhou Wang and Shansong Liu and Lingrui Mei and Peng Li and Junjie Wang and Jianwei Yu and Guojian Pang and Xu Li and Zihao Wang and Xiaohuan Zhou and Lijun Yu and Emmanouil Benetos and Yong Chen and Chenghua Lin and Xie Chen and Gus Xia and Zhaoxiang Zhang and Chao Zhang and Wenhu Chen and Xinyu Zhou and Xipeng Qiu and Roger Dannenberg and Jiaheng Liu and Jian Yang and Wenhao Huang and Wei Xue and Xu Tan and Yike Guo },
  journal={arXiv preprint arXiv:2503.08638},
  year={ 2025 }
}