Yang Bo
5 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
CIVIC is a path-consistent compact visual inference framework that achieves genuine hardware efficiency in Vision-Language Models by maintaining contiguous sequence representations across all inference stages, significantly reducing memory and latency.
CIVIC is a path-consistent compact visual inference framework that achieves genuine hardware efficiency in Vision-Language Models by maintaining contiguous sequence representations across all inference stages, significantly reducing memory and latency.
The paper introduces X-Stream, a new benchmark for multi-stream video understanding, and finds that current state-of-the-art MLLMs perform poorly when required to process multiple concurrent video streams.
The paper introduces X-Stream, a new benchmark for multi-stream video understanding, and finds that current state-of-the-art MLLMs perform poorly when required to process multiple concurrent video streams.
DMAD recasts distribution matching as classification and learns log-density ratios directly to improve few-step student training.
Papers
DMAD: Distribution Matching as Adversarial Distillation for Fast Visual Generation
Zhengming Yu, Junkun Yuan, Haotian Yang, Gordon Guocheng Qian +7 more
DMAD recasts distribution matching as classification and learns log-density ratios directly to improve few-step student training.