ImageNet
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ImageNet 100 (train)
0.44MAE
6
ImageNet-1k
0.975Precision @ 1% FPR
6
ImageNet 1K (val)
81.3Top-1 Accuracy
6
ImageNet and ImageNet-R (test)
8Cartoon Recall@10
6
ImageNet V2 (test)
1.14MAE (%)
6
ImageNet (train)
0.086Peak Training Memory (GB)
6
ImageNet Carnivore
13.65FID
6
ImageNet 100 (Patched) (val)
47Accuracy
6
ImageNet-100 Clean (val)
50.1Accuracy
6
ImageNet 512x512 (train)
3.85FID
6
ImageNet+
0.72RR
6
ImageNet 256x256x3 (train)
1,152.8Model Size Overhead (MB)
6
Tiny ImageNet (test)
24.5Accuracy
6
ImageNet 128 x 128 (train val)
1.81FID
6
ImageNet 64x64 down-sampled (train)
15.93IS
6
Tiny-ImageNet 5000 labels
44.18Accuracy
6
ImageNet 128x128 (test)
23.4FID
6
Tiny ImageNet 10%
51.18FID
6
Tiny ImageNet 50%
22.66FID
6
Tiny ImageNet 100%
14.84FID
6
TinyImageNet
99.7Fooling Rate
6
ImageNet-pair
93BA
6
ImageNet (val)
81.6Top-1 Acc
6
ImageNet
2Failure Rate
6
Imagenet
0Failure Rate
6