Image Classification on Contrastive Learning Pre-training No Attack (test)
61.3AccuracyContrastiveCrop
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ContrastiveCropCropping Strategy=Semantic-aware localization mechanism2022.11 | 61.3 | 40 | |
| No Defense (MoCo-v2)Data Augmentation Settings=Default MoCo-v2 operations2022.11 | 60.8 | 40 | |
| CompRess DistillationCompression Ratio=20%, Extra clean pre-training dataset=true2022.11 | 58.2 | 90 | |
| Localized CroppingCropping Strategy=Localized (delta = 0.2)2022.11 | 56.2 | 90 | |
| CompRess DistillationCompression Ratio=5%, Extra clean pre-training dataset=true2022.11 | 49.5 | 90 | |
| No Other Data AugsData Augmentation Settings=Only random cropping used2022.11 | 44.2 | 30 | |
| No Random CroppingCropping Strategy=None2022.11 | 32.4 | 220 |