Image Classification on Places LT 2 (test)
36.4Top-1 Acc (Overall)Feature Space Augmentation
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Feature Space AugmentationBackbone=ResNet-152, Pre-training=Pre-trained, Loss=Cross-Entropy2020.08 | 36.4 | 42.8 | 37.5 | 22.7 | |
| Feature Space AugmentationBackbone=ResNet-152, Pre-training=Pre-trained, Loss=Focal Loss2020.08 | 36 | 42.2 | 36.4 | 24 | |
| OLTRBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 35.9 | 44.7 | 37 | 25.3 | |
| Lifted LossBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 35.2 | 41.1 | 35.4 | 24 | |
| Range LossBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 35.1 | 41.1 | 35.4 | 23.2 | |
| FSLWFBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 34.9 | 43.9 | 29.9 | 29.5 | |
| FLBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 34.6 | 41.1 | 34.8 | 22.4 | |
| Plain ModelBackbone=ResNet-152, Pre-training=Pre-trained2020.08 | 27.2 | 45.9 | 22.4 | 0.36 |