Image Classification on CIFAR-100 SSL split 2500 labels (train test)
25.52Error RateCustomKD
Evaluation results
| Method | Links | |
|---|---|---|
| CustomKDCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 25.52 | |
| DKDCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 30.43 | |
| FitNetCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 30.87 | |
| Soft TargetCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 31.73 | |
| SimMatchCategory=SSL2025.03 | 32.54 | |
| FreeMatchCategory=SSL2025.03 | 32.79 | |
| SoftMatchCategory=SSL2025.03 | 33.05 | |
| LogitsCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 33.42 | |
| CCCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 33.72 | |
| RKDCategory=KD, Backbone=WideResNet28-2, Pre-trained with=AdaMatch2025.03 | 34.24 | |
| FixMatchCategory=SSL2025.03 | 34.29 | |
| AdaMatch*Category=SSL, Note=reproduced results using codes of USB benchmark2025.03 | 37.92 |