A Large-scale Study of Representation Learning with the Visual Task Adaptation Benchmark

Open paper
Benchmark Model Rank Results
image-classification-on-vtab-1kS4L-Exemplar-ResNet50-LargeHyperSweep#5Top-1 Accuracy: 72.7
image-classification-on-vtab-1kS4L-Rotation-ResNet50-LargeHyperSweep#6Top-1 Accuracy: 71.5
image-classification-on-vtab-1kImageNet-ResNet50-LargeHyperSweep#7Top-1 Accuracy: 71.2
image-classification-on-vtab-1kS4L-Rotation-ResNet50#9Top-1 Accuracy: 67.5
image-classification-on-vtab-1kS4L-Exemplar-ResNet50#10Top-1 Accuracy: 67.0
image-classification-on-vtab-1kImageNet-ResNet50#12Top-1 Accuracy: 65.6
image-classification-on-vtab-1kS4L-10%-Rotation-ResNet50#13Top-1 Accuracy: 64.8
image-classification-on-vtab-1kS4L-10%-Exemplar-ResNet50#14Top-1 Accuracy: 63.9
image-classification-on-vtab-1kImageNet-10%-ResNet50#15Top-1 Accuracy: 61.6
image-classification-on-vtab-1kSelfSup-Rotation-ResNet50#16Top-1 Accuracy: 59.5
image-classification-on-vtab-1kResNet50-LargeHyperSweep#17Top-1 Accuracy: 59.2
image-classification-on-vtab-1kBigBiGAN-ResNet50#18Top-1 Accuracy: 59.1
image-classification-on-vtab-1kSelfSup-Exemplar-ResNet50#20Top-1 Accuracy: 57.5
image-classification-on-vtab-1kSelfSup-Jigsaw-ResNet50#22Top-1 Accuracy: 51.1
image-classification-on-vtab-1kSelfSup-RelativePatchLoc-ResNet50#23Top-1 Accuracy: 50.8
image-classification-on-vtab-1kUnconditional-BigGAN-ResNet50#24Top-1 Accuracy: 44.0
image-classification-on-vtab-1kResNet50#25Top-1 Accuracy: 42.1
image-classification-on-vtab-1kVAE#26Top-1 Accuracy: 37.5
image-classification-on-vtab-1kWAE-MMD#27Top-1 Accuracy: 37.3
image-classification-on-vtab-1kConditional-BigGAN#28Top-1 Accuracy: 35.3
image-classification-on-vtab-1kWAE-GAN#29Top-1 Accuracy: 32.0
image-classification-on-vtab-1kWAE-UKL#30Top-1 Accuracy: 31.0

Rank counts only results with a code link.