MIT-States
Benchmarks
TaskDatasetSOTA resultTrendResults
MIT-States open world
37.2HM
38
MIT-States Closed World
0.413Harmonic Mean (HM)
32
MIT-States Closed World (test)
23.5AUC
27
MIT-States open-vocabulary
28.9HM
16
MIT-States Open-world (test)
50.7Seen Accuracy
14
MIT-states
52.5Top-1 Acc
12
MIT-States (test)
6.5AUC (1)
12
MIT-States (test)
19.9Top-1 Acc
11
MIT-States (val)
13.9R@1
9
MIT-States (test)
14.8Accuracy (Closed Set)
8
MIT-States 11 (test)
38.6Seen Accuracy
8
MIT-States
0.115AUC
7
MIT-States
15Rank-1 Accuracy
7
MIT-States (val)
4.3AUC (1)
6
MIT-States (compositional)
15.2Accuracy
6
MIT-States Filtered to 97 classes (test)
63.5Precision
5
MIT-states
45.5Top-1 Acc
4
MIT-States
—Accuracy (closed)
0