Image Classification on CIFAR-10 (test) (Linear/5-NN Accuracy)
92.47Linear Top-1 AccCW-RGP 4
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CW-RGP 4Backbone=ResNet-18, Positive views (s)=42022.10 | 92.47 | 90.74 | |
| W-MSE 4Backbone=ResNet-18, Positive views (s)=42022.10 | 91.99 | 89.87 | |
| CW-RGP 2Backbone=ResNet-18, Positive views (s)=22022.10 | 91.92 | 89.54 | |
| SimCLRBackbone=ResNet-182022.10 | 91.8 | 88.42 | |
| BYOLBackbone=ResNet-182022.10 | 91.73 | 89.45 | |
| W-MSE 2Backbone=ResNet-18, Positive views (s)=22022.10 | 91.55 | 89.69 | |
| SimSiamBackbone=ResNet-18, Reproduced=true2022.10 | 90.51 | 86.82 | |
| Shuffled-DBNBackbone=ResNet-18, Reproduced=true2022.10 | 90.45 | 88.15 | |
| VICRegBackbone=ResNet-18, Reproduced=true2022.10 | 90.32 | 88.41 | |
| Barlow TwinsBackbone=ResNet-18, Reproduced=true2022.10 | 88.51 | 86.53 | |
| Zero-ICLBackbone=ResNet-18, Reproduced=true2022.10 | 88.12 | 86.64 |