Recent advances in large language models (LLMs) and audio language models have significantly improved music generation, particularly in lyrics-to-song generation. However, existing approaches still struggle with the complex composition of songs and the scarcity of high-quality data, leading to limitations in sound quality, musicality, instruction following, and vocal-instrument harmony. To address these challenges, we introduce LeVo, an LM-based framework consisting of LeLM and a music codec. LeLM is capable of parallelly modeling two types of tokens: mixed tokens, which represent the combined audio of vocals and accompaniment to achieve vocal-instrument harmony, and dual-track tokens, which separately encode vocals and accompaniment for high-quality song generation. It employs two decoder-only transformers and a modular extension training strategy to prevent interference between different token types. To further enhance musicality and instruction following, we introduce a multi-preference alignment method based on Direct Preference Optimization (DPO). This method handles diverse human preferences through a semi-automatic data construction process and DPO post-training. Experimental results demonstrate that LeVo consistently outperforms existing methods on both objective and subjective metrics. Ablation studies further justify the effectiveness of our designs. Audio examples are available at this https URL.
View on arXiv@article{lei2025_2506.07520,
title={ LeVo: High-Quality Song Generation with Multi-Preference Alignment },
author={ Shun Lei and Yaoxun Xu and Zhiwei Lin and Huaicheng Zhang and Wei Tan and Hangting Chen and Jianwei Yu and Yixuan Zhang and Chenyu Yang and Haina Zhu and Shuai Wang and Zhiyong Wu and Dong Yu },
journal={arXiv preprint arXiv:2506.07520},
year={ 2025 }
}