Classification on Stanford Cars 30% labels (test)
87.7AccuracySelf-Tuning+HCR
Evaluation Results
| Method | Links | |
|---|---|---|
| Self-Tuning+HCRBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 87.7 | |
| Self-TuningBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 85.87 | |
| FixMatchBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 77.54 | |
| Pseudo-LabelingBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 67.02 | |
| Mean TeacherBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 66.02 | |
| UDABackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 64.16 | |
| SimCLRv2Backbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 61.7 | |
| Pi-modelBackbone=ResNet-50, Pre-trained=ImageNet, Label Proportion=30%2022.06 | 57.29 |