Elysium: Exploring Object-level Perception in Videos via MLLM

Benchmark Model Rank Results
zeroshot-video-question-answer-on-msrvtt-qaElysium#3Accuracy: 67.5Confidence Score: 3.2
zeroshot-video-question-answer-on-msvd-qaElysium#10Accuracy: 75.8Confidence Score: 3.7
zeroshot-video-question-answer-on-tgif-qaElysium#12Accuracy: 66.6Confidence Score: 3.6
zeroshot-video-question-answer-on-activitynet-qaElysium#24Accuracy: 43.4Confidence Score: 2.9