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

15 datasets

Benchmarks

Task NameDataset NameSOTA ResultTrend
Conformal Prediction15 datasets (average)
Coverage95.7
39
Model Selection15 datasets (average)
Identification Accuracy (8% Budget)83.5
12
Image Classification15 datasets Average (test)
Average Robust Accuracy45.39
12
Zero-shot Image Classification15 datasets Zero-shot (test)
Robust Accuracy (PGD-20, eps=1/255)44.34
5
Showing 4 of 4 rows