AVC-Reward: Event-Centric Audio-Video Consistency Evaluation
音视频一致性评估指标研究，非生成工具，无关
As joint audio-video generation models become more capable, they can generate audio-visual scenes exhibiting three forms of complexity: multi-event composition, multi-layer audio mixture, and asymmetric cross-modal relations. However, current evaluations largely reuse metrics fro
判断口径 agnes-3.0-flash · 入库 #94 · 回首页