Generalized Parametric Contrastive Learning

Open paper
Benchmark Model Rank Results
domain-generalization-on-imagenet-cGPaCo (ViT-L)#13mean Corruption Error (mCE): 39.0
domain-generalization-on-imagenet-rGPaCo (ViT-L)#14Top-1 Error Rate: 39.7
domain-generalization-on-imagenet-sketchGPaCo (ViT-L)#11Top-1 accuracy: 48.3
image-classification-on-imagenetGPaCo (ViT-L)#171Top 1 Accuracy: 86.01%
image-classification-on-imagenetGPaCo (Vit-B)#357Top 1 Accuracy: 84.0%
image-classification-on-imagenetGPaCo (ResNet-50)#706Top 1 Accuracy: 79.7%
image-classification-on-inaturalist-2018GPaCo (ResNet-152)#18Top-1 Accuracy: 78.1%
image-classification-on-inaturalist-2018GPaCo (ResNet-50)#22Top-1 Accuracy: 75.4%
long-tail-learning-on-imagenet-ltGPaCo (2-ResNeXt101-32x4d)#10Top-1 Accuracy: 63.2
long-tail-learning-on-inaturalist-2018GPaCo (2-R152)#6Top-1 Accuracy: 79.8%
long-tail-learning-on-inaturalist-2018GPaCo (ResNet-152)#7Top-1 Accuracy: 78.1%
long-tail-learning-on-inaturalist-2018GPaCo (ResNet-50)#11Top-1 Accuracy: 75.4%
long-tail-learning-on-places-ltGPaCo (ResNet-152)#11Top-1 Accuracy: 41.7
semantic-segmentation-on-ade20kGPaCo (Swin-L)#62Validation mIoU: 54.3
semantic-segmentation-on-pascal-contextGPaCo (ResNet101)#21mIoU: 56.2

Rank counts only results with a code link.