Streaming Video Instruction Tuning

Benchmark Model Rank Results
video-question-answering-on-mvbenchStreamo#22Avg.: 72.3
video-question-answering-on-ovo-benchStreamo-3B–Real-Time Visual Perception: 61.51Backward Tracing: 41.76…
video-question-answering-on-ovo-benchStreamo-7B–Real-Time Visual Perception: 65.98Backward Tracing: 46.10…