Fine-grained Image Classification on iNaturalist 2018 (test)
88.7AccuracyMetaFormer-2
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MetaFormer-2Backbone=MetaFormer-2, Pre-training=ImageNet-21k, Image size=384, Meta method=Transformer2022.03 | 88.7 | — | — | — | — | |
| MetaFormer-2Backbone=MetaFormer-2, Pre-training=ImageNet-1k, Image size=384, Meta method=Transformer2022.03 | 86.8 | — | — | — | — | |
| MetaFormer-1Backbone=MetaFormer-1, Pre-training=ImageNet-1k, Image size=384, Meta method=Transformer2022.03 | 86.5 | — | — | — | — | |
| MetaFormer-0Backbone=MetaFormer-0, Pre-training=ImageNet-1k, Image size=384, Meta method=Transformer2022.03 | 85.4 | — | — | — | — | |
| MetaFormer-2Backbone=MetaFormer-2, Pre-training=ImageNet-21k, Image size=384, Meta method=Image-Only2022.03 | 84.3 | — | — | — | — | |
| MetaFormer-2Backbone=MetaFormer-2, Pre-training=ImageNet-1k, Image size=384, Meta method=Image-Only2022.03 | 82.6 | — | — | — | — | |
| MetaFormer-1Backbone=MetaFormer-1, Pre-training=ImageNet-1k, Image size=384, Meta method=Image-Only2022.03 | 81.9 | — | — | — | — | |
| MetaFormer-0Backbone=MetaFormer-0, Pre-training=ImageNet-1k, Image size=384, Meta method=Image-Only2022.03 | 79.5 | — | — | — | — | |
| Presence-OnlyBackbone=Inception V3, Pre-training=ImageNet-1k, Image size=520, Meta method=Prior2022.03 | 77.5 | — | — | — | — | |
| Presence-OnlyBackbone=Inception V3, Pre-training=ImageNet-1k, Image size=299, Meta method=Prior2022.03 | 72.7 | — | — | — | — | |
| Presence-OnlyBackbone=Inception V3, Pre-training=ImageNet-1k, Image size=520, Meta method=Image-Only2022.03 | 66.2 | — | — | — | — | |
| Presence-OnlyBackbone=Inception V3, Pre-training=ImageNet-1k, Image size=299, Meta method=Image-Only2022.03 | 60.2 | — | — | — | — | |
| ImgGeo-prior model integration=None (Image only)2025.02 | — | 66.1 | 83.3 | 88 | 92.2 | |
| Img+CSPGeo-prior model integration=CSP2025.02 | — | 72.9 | 87.9 | 91.6 | 94.8 | |
| Img+CSP_INat*Geo-prior model integration=CSP_INat*2025.02 | — | 74.4 | 88.8 | 92.2 | 94.9 | |
| Img+GeoCLIPGeo-prior model integration=GeoCLIP2025.02 | — | 72.9 | 88.2 | 91.9 | 95.2 | |
| Img+RANGEGeo-prior model integration=RANGE2025.02 | — | 75.2 | 89.6 | 92.9 | 95.5 | |
| Img+RANGE+Geo-prior model integration=RANGE+2025.02 | — | 75.1 | 89.5 | 92.8 | 95.5 | |
| Img+SatCLIPGeo-prior model integration=SatCLIP2025.02 | — | 75.1 | 88.7 | 91.9 | 94.5 |