Multi-label image classification on VG-500 (test)
43.1mAPML-Decoder + AAM
Evaluation Results
| Method | Links | |
|---|---|---|
| ML-Decoder + AAMBackbone=TResNet-L, Input resolution=576x576, GFLOPS=59.632022.09 | 43.1 | |
| Q2LBackbone=TResNet-L, Input resolution=512x512, GFLOPS=119.372022.09 | 42.5 | |
| ML-Decoder + AAMBackbone=EfficientNet-V2-s, Input resolution=576x576, GFLOPS=20.162022.09 | 42 | |
| ML-DecoderBackbone=EfficientNet-V2-s, Input resolution=576x576, GFLOPS=20.16, Training strategy=Trained by current paper2022.09 | 41.2 | |
| Q2LBackbone=EfficientNet-V2-s, Input resolution=576x576, GFLOPS=32.81, Training strategy=Trained by current paper2022.09 | 40.35 | |
| ASLBackbone=EfficientNet-V2-s, Input resolution=576x576, GFLOPS=17.9, Training strategy=Trained by current paper2022.09 | 38.84 | |
| C-TranBackbone=ResNet101, Input resolution=576x5762022.09 | 38.4 | |
| SSGRLbackbone=ResNet-1012019.08 | 36.6 | |
| ResNet-SRNbackbone=ResNet-1012019.08 | 33.5 | |
| ResNet-101backbone=ResNet-1012019.08 | 30.9 |