Multi-label classification on LVIS v1.0 (test)
78.57mAP (C)P-ASL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| P-ASLMode=Selective, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 78.57 | 96.8 | 99.81 | |
| ASLMode=Negative, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 78.32 | 96.77 | 98.43 | |
| CEMode=Negative, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 77.82 | 96.66 | 97.2 | |
| SEMode=Negative, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 77.81 | 96.6 | 97.28 | |
| CEMode=Ignore, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 74.49 | 95.7 | 99.81 | |
| wCEMode=Ignore, Backbone=ResNet-101, Pre-trained=ImageNet2021.10 | 74.15 | 95.2 | 99.8 |