Image Classification on ImageNet 100B
95.4Top-5 AccOurs
Evaluation results
| Method | Links | |
|---|---|---|
| OursDistortion Type=Random missing pixels (RM), Evaluation Protocol=Linear probing, Pre-trained on=ImageNet-1002021.10 | 95.4 | |
| OursDistortion Type=Additive Gaussian noise (GN), Evaluation Protocol=Linear probing, Pre-trained on=ImageNet-1002021.10 | 94.74 | |
| BaselineDistortion Type=Additive Gaussian noise (GN), Evaluation Protocol=Linear probing, Pre-trained on=ImageNet-1002021.10 | 89.81 | |
| BaselineDistortion Type=Random missing pixels (RM), Evaluation Protocol=Linear probing, Pre-trained on=ImageNet-1002021.10 | 88.35 |