1 paper · 1 filter
Kaiwen Wang, Kaili Zheng, Rongrong Deng +3
Recent multimodal large language models (MLLMs) have shown strong capabilities in general video understanding, driving growing interest in automatic sports commentary generation. H…