Multi-label classification on VG500
42.5mAPQ2L-TResL-22k
Evaluation results
| Method | Links | |
|---|---|---|
| Q2L-TResL-22kBackbone=TResNetL, Pre-training=ImageNet-22k, Resolution=512 x 5122021.07 | 42.5 | |
| Q2L-R101Backbone=ResNet-101, Resolution=512 x 5122021.07 | 39.5 | |
| C-TranBackbone=ResNet-101, Resolution=576 x 5762021.07 | 38.4 | |
| SSGRLBackbone=ResNet-101, Resolution=576 x 5762021.07 | 36.6 | |
| ResNet-SRNBackbone=ResNet-101, Resolution=576 x 5762021.07 | 33.5 | |
| ResNet-101Backbone=ResNet-101, Resolution=576 x 5762021.07 | 30.9 |