Classification Datasets
Benchmarks
TaskDatasetSOTA resultTrendResults
11 Classification Datasets (ImageNet, SUN397, FGVC Aircraft, EuroSAT, Stanford Cars, Food-101, Oxford-IIIT Pets, Oxford Flowers 102, Caltech101, DTD, UCF101)
68.4ImageNet Accuracy
39
4 classification datasets average
99.8RAcc
35
53 classification datasets (unseen)
75.42Mean Accuracy
18
Classification Datasets (MMLU, OBQA, ARC-e, WinoGrande, ARC-c, PIQA, HellaSwag)
37.1MMLU (5-shot)
18
80 classification datasets
0.11Median Effect Size (F1 pts)
17
11 classification datasets (test)
76.77ImageNet Accuracy
16
medium-sized classification datasets
78.58Accuracy
14
Classification Datasets Average (test)
72.5NAURC
12
Classification Datasets
100Accuracy
10
25 Classification Datasets
89.1Mean Accuracy
10
7 classification datasets (Iris, Wine, Breast Cancer, Digits, etc.) (cross-validation)
91.17Accuracy
10
50 classification datasets
84.36Mean Accuracy
10
6 out-of-domain classification datasets (test)
65.2Accuracy
9
Classification Datasets
0.05Avg. JSD
2
15 Classification Datasets
1,067TabMixNN Wins
1