4 papers
MoE-TTS: Enhancing Out-of-Domain Text Understanding for Description-based TTS via Mixture-of-Experts
Heyang Xue, Xuchen Song, Yu Tang +4
Description-based text-to-speech (TTS) models exhibit strong performance on in-domain text descriptions, i.e., those encountered during training. However, in real-world application…
SkyReels-V2: Infinite-length Film Generative Model
Guibin Chen, Dixuan Lin, Jiangping Yang +22
Recent advances in video generation have been driven by diffusion models and autoregressive frameworks, yet critical challenges persist in harmonizing prompt adherence, visual qual…
SkyReels-A2: Compose Anything in Video Diffusion Transformers
Zhengcong Fei, Debang Li, Di Qiu +8
This paper presents SkyReels-A2, a controllable video generation framework capable of assembling arbitrary visual elements (e.g., characters, objects, backgrounds) into synthesized…
Analyzable Chain-of-Musical-Thought Prompting for High-Fidelity Music Generation
Max W. Y. Lam, Yijin Xing, Weiya You +14
Autoregressive (AR) models have demonstrated impressive capabilities in generating high-fidelity music. However, the conventional next-token prediction paradigm in AR models does n…