Image Classification on CIFAR-100 ImageNet-1K transfer (test)
0.6133SAStandard CL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Standard CLProtocol=AFF, Pre-training=Standard Contrastive Learning, Running time (hours)=147.42023.02 | 0.6133 | 0.2511 | |
| ACL with RCSProtocol=AFF, Pre-training=Robustness-Aware Coreset Selection, Running time (hours)=111.82023.02 | 0.6025 | 0.2824 | |
| ACL with RandomProtocol=AFF, Pre-training=Adversarial Contrastive Learning, Running time (hours)=94.32023.02 | 0.5963 | 0.2546 | |
| Standard CLProtocol=SLF, Pre-training=Standard Contrastive Learning, Running time (hours)=147.42023.02 | 0.5734 | 0.0001 | |
| ACL with RCSProtocol=SLF, Pre-training=Robustness-Aware Coreset Selection, Running time (hours)=111.82023.02 | 0.4028 | 0.1455 | |
| ACL with RandomProtocol=SLF, Pre-training=Adversarial Contrastive Learning, Running time (hours)=94.32023.02 | 0.3853 | 0.105 | |
| ACL with RCSProtocol=ALF, Pre-training=Robustness-Aware Coreset Selection, Running time (hours)=111.82023.02 | 0.3315 | 0.1489 | |
| ACL with RandomProtocol=ALF, Pre-training=Adversarial Contrastive Learning, Running time (hours)=94.32023.02 | 0.2844 | 0.1193 | |
| Standard CLProtocol=ALF, Pre-training=Standard Contrastive Learning, Running time (hours)=147.42023.02 | 0.0932 | 0.0006 |