Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

fMoW

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image ClassificationfMoW (test)
Top-1 Accuracy98.96
60
Image ClassificationfMoW (val)
Accuracy79.3
34
Image ClassificationfMoW
Accuracy81.7
21
Multi-label ClassificationFMoW (test)
WG Score27.75
20
Remote Sensing ClassificationfMoW
Top-1 Accuracy56.56
15
Image ClassificationFMOW ID
Accuracy68.6
15
Satellite image generationFMoW
FID32.11
13
Post-deployment performance monitoringFMoW Temporal Shift
R2 Score0.869
12
AI-generated image detectionFMoW (test)
F1 Score (Fake Class)99.88
12
Image ClassificationfMoW-RGB
Top-1 Acc69.2
12
Satellite Image ClassificationFMoW WILDS
Worst-Region Accuracy42.8
11
Image ClassificationFMoW OOD (test)
Worst Group Accuracy42.8
10
Image ClassificationFMoW ID (test)
Accuracy69.2
10
Active LearningfMoW (test)
Accuracy98.41
10
Image ClassificationFMoW WILDS (test)
WC Accuracy46.4
10
Image ClassificationFMoW-WILDS out-of-distribution (val)
Accuracy59.2
10
Image Classification CalibrationFMoW
Accuracy53.81
9
Scene ClassificationfMoW-S2 (test)
Top-1 Accuracy64.38
9
Infrastructure classificationFMOW WILDS (test)
Average Accuracy74.8
9
ClassificationfMoW-RGB (val)
Top-1 Accuracy78.14
8
Image ClassificationfMoW-Sentinel (val)
Top-1 Accuracy63.23
8
Temporal ClassificationfMoW (test)
F1 Score70.11
8
Domain transfer difficulty predictionFMoW-Wilds
Absolute Pearson Correlation (|rho|)0.65
7
Satellite Image ClassificationFMoW OOD WILDS (test)
Worst-Region Accuracy34.6
7
Image ClassificationFMOW-WILDS
WG Error6.8
7
Showing 25 of 37 rows