Just Ask: Learning to Answer Questions from Millions of Narrated Videos

Benchmark Model Rank Results
video-question-answering-on-activitynet-qaJust Ask (fine-tune)#22Accuracy: 38.9
video-question-answering-on-activitynet-qaJust Ask (0-shot)#32Accuracy: 12.2
video-question-answering-on-how2qaJust Ask#3Accuracy: 84.4
video-question-answering-on-how2qaJust Ask (0-shot)#7Accuracy: 51.1
video-question-answering-on-ivqaJust Ask (fine-tune)#3Accuracy: 35.4
video-question-answering-on-ivqaJust Ask (0-shot)#5Accuracy: 12.2
visual-question-answering-on-msrvtt-qa-2Just Ask#2Accuracy: 0.415