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

Robustness Benchmark

Benchmarks

Task NameDataset NameSOTA ResultTrend
Zero-shot Image Classification14 Robustness Benchmark Datasets (ImageNet, CalTech, Cars, CIFAR10, CIFAR100, DTD, EuroSAT, FGVC, Flowers, ImageNet-R, ImageNet-S, PCAM, OxfordPets, STL-10) (test)
ImageNet Accuracy80.11
16
Mobile UI ControlRobustness Benchmark
LR43.3
10
Robustness EvaluationRobustness benchmark
Indirect PI Accuracy100
4
Showing 3 of 3 rows