Unsupervised Image Classification on ImageNet Superclasses 128 classes WordNet hierarchy (val)
0.718AccuracySelf-Classifier
Evaluation results
| Method | Links | |
|---|---|---|
| Self-ClassifierBackbone=ResNet-502021.03 | 0.718 | |
| SCANBackbone=ResNet-50, Pre-training=ImageNet-pretrained2021.03 | 0.712 | |
| OBoWBackbone=ResNet-50, Evaluation Protocol=k-means classifier2021.03 | 0.674 | |
| MoCoV2Backbone=ResNet-50, Evaluation Protocol=k-means classifier2021.03 | 0.638 | |
| BarlowTBackbone=ResNet-50, Evaluation Protocol=k-means classifier2021.03 | 0.627 | |
| DINOBackbone=ResNet-50, Evaluation Protocol=k-means classifier2021.03 | 0.607 | |
| SwAVBackbone=ResNet-50, Evaluation Protocol=k-means classifier2021.03 | 0.58 | |
| SeLaBackbone=ResNet-50, Pre-training=ImageNet-pretrained2021.03 | 0.406 |