Image Classification on ImageNetV2 Threshold 0.7 (test)
74.5Top-1 AccRELICv2
Evaluation Results
| Method | Links | |
|---|---|---|
| RELICv2Protocol=Linear evaluation, Backbone=ResNet-50, Pre-training=ImageNet, Evaluation=Zero-shot2022.01 | 74.5 | |
| SupervisedProtocol=Linear evaluation, Backbone=ResNet-50, Pre-training=ImageNet, Evaluation=Zero-shot2022.01 | 73.9 | |
| RELICProtocol=Linear evaluation, Backbone=ResNet-50, Pre-training=ImageNet, Evaluation=Zero-shot2022.01 | 72.3 | |
| BYOLProtocol=Linear evaluation, Backbone=ResNet-50, Pre-training=ImageNet, Evaluation=Zero-shot2022.01 | 71.6 | |
| RELICv2Epochs=3000, Pre-training Dataset=JFT-300M, Evaluation Protocol=linear evaluation2022.01 | 67.5 | |
| RELICv2Epochs=5000, Pre-training Dataset=JFT-300M, Evaluation Protocol=linear evaluation2022.01 | 67.3 | |
| RELICv2Epochs=1000, Pre-training Dataset=JFT-300M, Evaluation Protocol=linear evaluation2022.01 | 66.7 | |
| SimCLRProtocol=Linear evaluation, Backbone=ResNet-50, Pre-training=ImageNet, Evaluation=Zero-shot2022.01 | 61.7 |