OmniAgent: Audio-Guided Active Perception Agent for Omnimodal Audio-Video Understanding

Benchmark Model Rank Results
audio-visual-question-answering-on-dailyomniOmniAgent#1Accuracy: 82.71
audio-visual-question-answering-on-omnivideobenchOmniAgent#1Accuracy: 59.1
audio-visual-question-answering-on-worldsenseOmniAgent#1Accuracy: 61.2