Semi-supervised Image Classification on ImageNet 10% labeled samples (train test)
91.4Top-5 AccUDA+AutoDropout
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UDA+AutoDropoutBackbone=ResNet-502021.01 | 91.4 | 72.9 | |
| FixMatchBackbone=ResNet-502021.01 | 89.1 | 71.5 | |
| SimCLRBackbone=ResNet-502021.01 | 89 | 69.3 | |
| UDABackbone=ResNet-502021.01 | 88.8 | 68.8 | |
| LLPBackbone=ResNet-502021.01 | 88.5 | — |