Benchmarking Self-Supervised Learning on Diverse Pathology Datasets

Open paper
Benchmark Model Rank Results
classification-on-mhistMoCo-v2 (ResNet-50)#2Accuracy: 85.88
classification-on-mhistSupervised (ViT-S/16)#5Accuracy: 81.68
classification-on-mhistBarlow Rwins (ResNet-50)#6Accuracy: 81.27
classification-on-mhistDINO (ViT-S/16)#7Accuracy: 79.43
classification-on-mhistSupervised (ResNet-50)#8Accuracy: 78.92
classification-on-mhistSwAV (ResNet-50)#9Accuracy: 77.99

Rank counts only results with a code link.