MuLTI: Efficient Video-and-Language Understanding with Text-Guided MultiWay-Sampler and Multiple Choice Modeling

Benchmark Model Rank Results
video-retrieval-on-didemoMuLTI–text-to-video R@1: 56.5text-to-video R@5: 80.2…
video-retrieval-on-msr-vtt-1kaMuLTI–text-to-video R@1: 54.7text-to-video R@5: 77.7…
visual-question-answering-on-msrvtt-qaMuLTI–Accuracy: 0.478
visual-question-answering-on-msvd-qaMuLTI–Accuracy: 0.547