Multi-label classification on OpenImages V3
73.19Performance Group 1P-ASL, Selective
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| P-ASL, SelectiveBackbone=ResNet-101, Pre-trained=ImageNet2021.10 | 73.19 | 78.61 | 85.11 | 87.7 | 90.61 | 83.03 | |
| IMCLBackbone=ResNet-101, Pre-trained=ImageNet2021.10 | 70.95 | 72.59 | 77.64 | 81.83 | 87.34 | 78.07 | |
| Curriculum LabelingBackbone=ResNet-101, Pre-trained=ImageNet2021.10 | 70.37 | 71.32 | 76.23 | 80.54 | 86.81 | 77.05 | |
| Latent Noise (visual)Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 69.37 | 70.41 | 74.79 | 79.2 | 85.51 | 75.86 | |
| CNN-RNNBackbone=ResNet-101, Pre-trained=ImageNet2021.10 | 68.76 | 69.7 | 74.18 | 78.52 | 84.61 | 75.16 |