5 papers
Shao: Scaling Acoustic Token Language Models Toward High-Fidelity Music Generation
Jiafeng Liu, Yuanliang Dong, Hongjia Liu +8
A common design pattern in high-quality music generation is to handle structure and fidelity in different representation spaces: a generator first models high-level structure, foll…
DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing
Dianyi Wang, Ruihang Li, Feng Han +17
Current unified multimodal models for image generation and editing typically rely on massive parameter scales (e.g., >10B), entailing prohibitive training costs and deployment foot…
TISDiSS: A Training-Time and Inference-Time Scalable Framework for Discriminative Source Separation
Yongsheng Feng, Yuetonghui Xu, Jiehui Luo +4
Source separation is a fundamental task in speech, music, and audio processing, and it also provides cleaner and larger data for training generative models. However, improving sepa…
A Training-free LLM Framework with Interaction between Contextually Related Subtasks in Solving Complex Tasks
Hongjia Liu, Jinlong Li
Large language models (LLMs) have shown remarkable capabilities in solving complex tasks. Recent work has explored decomposing such tasks into subtasks with independent contexts. H…
SCott: Accelerating Diffusion Models with Stochastic Consistency Distillation
Hongjian Liu, Qingsong Xie, TianXiang Ye +6
The iterative sampling procedure employed by diffusion models (DMs) often leads to significant inference latency. To address this, we propose Stochastic Consistency Distillation (S…