MSR-VTT
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MSR-VTT (VerbH)
85.2Accuracy
7
MSR-VTT 7K
46.9Recall@1
7
MSR-VTT 7K (train)
43.9R@1
7
MSR-VTT
3,751Initial Memory (MB)
6
MSR-VTT
14.4R1
6
MSR-VTT
91.3FVD
6
MSR-VTT full 7k train (test)
34.9R@1
6
MSR-VTT 1K (val)
0.562R@1
6
MSR-VTT (train)
0.267SSLMem (T)
5
MSR-VTT
45.2Accuracy (50% Frames)
5
MSR-VTT English
13.3BLEU@4
5
MSR-VTT (split1)
28.4R@1
5
MSR-VTT (val)
41.7CIDEr
5
MSR-VTT unseen adverb-action compositions (test)
61Accuracy
5
MSR-VTT
99.99SB
5
MSR-VTT 9 (test)
8.71R@1
4
MSR-VTT
24.1Recall@1
4
MSR-VTT 3 modal
26.8BLEU@1
4
MSR-VTT (test-1k)
3.02S Score
4
MSR-VTT (val)
0.301Frame Consistency (F.C.)
4
MSR-VTT
40.9Top-1 Accuracy
4
MSR-VTT
42Top-1 Accuracy
4
MSR-VTT (standard)
55.2R@1
4
MSR-VTT (test)
33.1R@1
4
MSR-VTT 1 caption per video
11.2Recall@1
4