Image Classification on iNaturalist 2019 (val)
83.2Top-1 AccuracyTNT-B
Evaluation results
| Method | Links | |
|---|---|---|
| TNT-BParams (M)=65.6, Resolution=384x384, Pre-training=ImageNet2021.02 | 83.2 | |
| TNT-SParams (M)=23.8, Resolution=384x384, Pre-training=ImageNet2021.02 | 81.4 | |
| Grafit RegNetY-8GFParams (M)=39.2, Resolution=384x384, Pre-training=ImageNet2021.02 | 80 | |
| iBOTBackbone=ViT-B/16, Evaluation Protocol=Fine-tuning2021.11 | 79.6 | |
| XCiT-L24/16Architecture=XCiT-L24/16, distillation=true2021.06 | 79.3 | |
| DeiT III-LBackbone=ViT-L, Pre-training=ImageNet-1k, Epochs=400, Resolution=224x224, Crop ratio=1.02022.04 | 79.3 | |
| BEITBackbone=ViT-B/16, Evaluation Protocol=Fine-tuning2021.11 | 79.2 | |
| DINOBackbone=ViT-B/16, Evaluation Protocol=Fine-tuning2021.11 | 78.6 | |
| iBOTBackbone=ViT-S/16, Evaluation Protocol=Fine-tuning2021.11 | 78.5 | |
| DeiT-B/16Architecture=DeiT-B/16, distillation=true2021.06 | 78.4 | |
| DINOBackbone=ViT-S/16, Evaluation Protocol=Fine-tuning2021.11 | 78.2 | |
| XCiT-M24/16Architecture=XCiT-M24/16, distillation=true2021.06 | 78.1 | |
| DeiT III-BBackbone=ViT-B, Pre-training=ImageNet-1k, Epochs=400, Resolution=224x224, Crop ratio=1.02022.04 | 78 | |
| Rand.Backbone=ViT-B/16, Evaluation Protocol=Fine-tuning2021.11 | 77.7 | |
| DeiT-BPre-training=ImageNet-1k, Crop ratio=0.8752022.04 | 77.7 | |
| Rand.Backbone=ViT-S/16, Evaluation Protocol=Fine-tuning2021.11 | 76.6 | |
| BEITBackbone=ViT-S/16, Evaluation Protocol=Fine-tuning2021.11 | 76.5 | |
| XCiT-S24/16Architecture=XCiT-S24/16, distillation=true2021.06 | 76.1 | |
| Grafit ResNet-50Params (M)=25.6, Resolution=384x384, Pre-training=ImageNet2021.02 | 75.9 | |
| Grafit ResNet-50 [52]Pre-training=ImageNet-1k, Crop ratio=0.8752022.04 | 75.9 | |
| DeiT III-SBackbone=ViT-S, Pre-training=ImageNet-1k, Epochs=400, Resolution=224x224, Crop ratio=1.02022.04 | 72.7 |