Long-tailed Image Classification on iNat (val test)
80.2Overall AccuracyRAC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RACBackbone=ViT-B/16, Linguistic data=true2024.10 | 80.2 | 75.9 | 80.5 | 81.1 | |
| VL-LTRBackbone=ViT-B/16, Linguistic data=true2024.10 | 76.8 | — | — | — | |
| GNM-PTBackbone=ViT-B/16, Visual-only=true, DRW=true2024.10 | 76.5 | 61.5 | 77.1 | 79.3 | |
| GNM-PTBackbone=ViT-B/16, Visual-only=true, DRW=false2024.10 | 76.3 | 76.3 | 77.6 | 75 | |
| LPTBackbone=ViT-B/16, Visual-only=true2024.10 | 76.1 | — | 79.3 | — | |
| LiVTBackbone=ViT-B/16, Visual-only=true2024.10 | 76.1 | 78.9 | 76.5 | 74.8 | |
| GPaCoBackbone=ResNet50, Type=DNN-based2024.10 | 75.4 | — | — | — | |
| SHIKEBackbone=ResNet50, Type=DNN-based2024.10 | 75.4 | — | — | — | |
| NCLBackbone=ResNet50, Type=DNN-based2024.10 | 74.9 | 72.7 | 75.6 | 74.5 | |
| RIDEBackbone=ResNet50, Type=DNN-based2024.10 | 72.8 | 76.5 | 74.2 | 70.5 | |
| GCLBackbone=ResNet50, Type=DNN-based2024.10 | 72 | — | — | — | |
| MisLASBackbone=ResNet50, Type=DNN-based2024.10 | 71.6 | 73.2 | 72.4 | 70.4 | |
| ImbSAMBackbone=ResNet50, SAM=true2024.10 | 71.1 | 68.2 | 72.5 | 72.9 | |
| CCSAMBackbone=ResNet50, SAM=true2024.10 | 70.9 | 65.4 | 70.9 | 72.2 | |
| LWSBackbone=ResNet50, Type=DNN-based2024.10 | 70.5 | 72.9 | 71.2 | 69.2 | |
| LDAM+SAMBackbone=ResNet50, SAM=true2024.10 | 70.1 | 64.1 | 70.5 | 71.2 | |
| DecoderBackbone=ViT-B/16, Visual-only=true2024.10 | 59.2 | — | — | — |