Image Classification on ImageNet-C Severity 5
45.19Top-1 AccuracyReference Masking Regularization
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Reference Masking RegularizationBackbone=SWINv2-B, Parameters=87.8M2024.06 | 45.19 | 58.77 | |
| Reference Masking RegularizationBackbone=VITb16, Parameters=86.3M2024.06 | 44.52 | 54.94 | |
| Reference Masking RegularizationBackbone=SWINv2-S, Parameters=49.7M2024.06 | 44.33 | 58.25 | |
| StandardBackbone=SWINv2-S, Parameters=49.7M2024.06 | 42.85 | 53.7 | |
| StandardBackbone=SWINv2-B, Parameters=87.8M2024.06 | 41.87 | 53.75 | |
| StandardBackbone=VITb16, Parameters=86.3M2024.06 | 41.09 | 50.07 | |
| Reference Masking RegularizationBackbone=SWINv2-T, Parameters=28.4M2024.06 | 38.58 | 51.76 | |
| StandardBackbone=SWINv2-T, Parameters=28.4M2024.06 | 35.66 | 46.54 |