Class-Balanced Loss Based on Effective Number of Samples

Open paper
Benchmark Model Rank Results
image-classification-on-inaturalist-2018ResNet-152#36Top-1 Accuracy: 69.05%
image-classification-on-inaturalist-2018ResNet-101#38Top-1 Accuracy: 67.98%
image-classification-on-inaturalist-2018ResNet-50#45Top-1 Accuracy: 64.16%
long-tail-learning-on-cifar-10-lt-r-10Class-balanced Focal Loss#37Error Rate: 12.90
long-tail-learning-on-cifar-10-lt-r-10Class-balanced Reweighting#40Error Rate: 13.46
long-tail-learning-on-cifar-100-lt-r-100Cross-Entropy (CE)#56Error Rate: 61.68
long-tail-learning-on-coco-mltCB Loss(ResNet-50)#7Average mAP: 49.06
long-tail-learning-on-voc-mltCB Focal(ResNet-50)#7Average mAP: 75.24

Rank counts only results with a code link.