Image Classification on ImageNet Real (test)
91.02Top-1 AccuracyMPL
Evaluation results
| Method | Links | |
|---|---|---|
| MPLBackbone=EfficientNet-L22021.06 | 91.02 | |
| PaLI-XResolution=756, Evaluation Protocol=fine-tuned2023.05 | 90.98 | |
| ViT-G/14Backbone=ViT-G/142021.06 | 90.81 | |
| ViT-H/14Backbone=ViT-H/142021.06 | 90.72 | |
| VOLO-D5↑448Architecture=VOLO, Params=296M, FLOPs=304B, Train size=224, Test size=4482021.06 | 90.6 | |
| VOLO-D5↑512Architecture=VOLO, Params=296M, FLOPs=412B, Train size=224, Test size=5122021.06 | 90.6 | |
| BiT-LBackbone=ResNet2021.06 | 90.54 | |
| VOLO-D4↑448Architecture=VOLO, Params=193M, FLOPs=197B, Train size=224, Test size=4482021.06 | 90.5 | |
| PaLI-XResolution=224, Evaluation Protocol=fine-tuned2023.05 | 90.36 | |
| CaiT-M36↑448 [52]Architecture=Transformer, Params=271M, FLOPs=248B, Train size=224, Test size=4482021.06 | 90.2 | |
| CaiT-M48↑448 [52]Architecture=Transformer, Params=356M, FLOPs=330B, Train size=224, Test size=4482021.06 | 90.2 | |
| Fix-EfficientNet-B8 [50, 53]Architecture=CNN, Params=87M, FLOPs=89.5B, Train size=672, Test size=8002021.06 | 90 | |
| VOLO-D3↑448Architecture=VOLO, Params=86M, FLOPs=67.9B, Train size=224, Test size=4482021.06 | 90 | |
| VOLO-D5Architecture=VOLO, Params=296M, FLOPs=69.0B, Train size=224, Test size=2242021.06 | 89.9 | |
| CaiT-S36 384 [52]Architecture=Transformer, Params=68M, FLOPs=48.0B, Train size=224, Test size=3842021.06 | 89.8 | |
| VOLO-D2↑384Architecture=VOLO, Params=59M, FLOPs=46.1B, Train size=224, Test size=3842021.06 | 89.7 | |
| VOLO-D4Architecture=VOLO, Params=193M, FLOPs=43.8B, Train size=224, Test size=2242021.06 | 89.7 | |
| VOLO-D1↑384Architecture=VOLO, Params=27M, FLOPs=22.8B, Train size=224, Test size=3842021.06 | 89.6 | |
| VOLO-D3Architecture=VOLO, Params=86M, FLOPs=20.6B, Train size=224, Test size=2242021.06 | 89.6 | |
| LV-ViT-M↑384 [32]Architecture=Transformer, Params=56M, FLOPs=42.2B, Train size=224, Test size=3842021.06 | 89.5 | |
| NFNet-F3 [2]Architecture=CNN, Params=255M, FLOPs=115.0B, Train size=320, Test size=4162021.06 | 89.4 | |
| NFNet-F4 [2]Architecture=CNN, Params=316M, FLOPs=215B, Train size=384, Test size=5122021.06 | 89.4 | |
| VOLO-D2Architecture=VOLO, Params=59M, FLOPs=14.1B, Train size=224, Test size=2242021.06 | 89.3 | |
| NFNet-F5 [2]Architecture=CNN, Params=377M, FLOPs=290B, Train size=416, Test size=5442021.06 | 89.2 | |
| NFNet-F6 [2]+SAMArchitecture=CNN, Params=438M, FLOPs=377B, Train size=448, Test size=5762021.06 | 89.2 | |
| VOLO-D1Architecture=VOLO, Params=27M, FLOPs=6.8B, Train size=224, Test size=2242021.06 | 89 | |
| LV-ViT-S↑384 [32]Architecture=Transformer, Params=26M, FLOPs=22.2B, Train size=224, Test size=3842021.06 | 88.9 | |
| NFNet-F1 [2]Architecture=CNN, Params=133M, FLOPs=35.5B, Train size=224, Test size=3202021.06 | 88.9 | |
| NFNet-F2 [2]Architecture=CNN, Params=194M, FLOPs=62.6B, Train size=256, Test size=3522021.06 | 88.9 | |
| PaLI-17BResolution=224, Evaluation Protocol=fine-tuned2023.05 | 88.84 | |
| EfficientNet-B5 [50]Architecture=CNN, Params=30M, FLOPs=9.9B, Train size=456, Test size=4562021.06 | 88.3 | |
| NFNet-F0 [2]Architecture=CNN, Params=72M, FLOPs=12.4B, Train size=192, Test size=2562021.06 | 88.1 | |
| T2T-ViT-14 384 [68]Architecture=Transformer, Params=22M, FLOPs=17.1B, Train size=224, Test size=3842021.06 | 87.8 | |
| EfficientNetV2-rw-MTraining procedure=A12021.10 | 87.1 | |
| FKDBackbone=ResNet-1012021.12 | 87 | |
| ECA-Resnet269-DTraining procedure=A12021.10 | 86.9 | |
| T2T-ViT-14 [68]Architecture=Transformer, Params=22M, FLOPs=5.2B, Train size=224, Test size=2242021.06 | 86.8 | |
| DeiT-B [51]Architecture=Transformer, Params=86M, FLOPs=17.5B, Train size=224, Test size=2242021.06 | 86.7 | |
| RegNetY-4GFTraining procedure=A12021.10 | 86.7 | |
| RegNetY-8GFTraining procedure=A12021.10 | 86.7 | |
| EfficientNet-B3Training procedure=A12021.10 | 86.7 | |
| RegNetY-32GFTraining procedure=A12021.10 | 86.6 | |
| ReLabelBackbone=ResNet-101, Pre-trained=true2021.12 | 86.5 | |
| ResNet-152Training procedure=A12021.10 | 86.4 | |
| RegNetY-16GFTraining procedure=A12021.10 | 86.4 | |
| ResNet-101Training procedure=A12021.10 | 86.3 | |
| ECA-ResNet50-TTraining procedure=A12021.10 | 86.1 | |
| SENet-154Training procedure=A12021.10 | 86 | |
| EfficientNet-B2Training procedure=A12021.10 | 86 | |
| ResNet-50-DTraining procedure=A12021.10 | 85.9 | |
| EfficientNet-B4Training procedure=A12021.10 | 85.9 | |
| SE-ResNet-50Training procedure=A12021.10 | 85.8 | |
| DeiT-S [51]Architecture=Transformer, Params=22M, FLOPs=4.6B, Train size=224, Test size=2242021.06 | 85.7 | |
| ResNet-50Training procedure=A12021.10 | 85.7 | |
| ViT-STraining procedure=A12021.10 | 85.6 | |
| ResNeXt-50-32x4dTraining procedure=A12021.10 | 85.5 | |
| FKDBackbone=ResNet-502021.12 | 85.5 | |
| EfficientNet-B1Training procedure=A12021.10 | 85.3 | |
| ReLabelBackbone=ResNet-502021.12 | 85 | |
| ViT-BTraining procedure=A12021.10 | 84.8 | |
| EfficientNetV2-rw-STraining procedure=A12021.10 | 84.8 | |
| EfficientNet-B0Training procedure=A12021.10 | 83.8 | |
| ViT-B/16 [14]Architecture=Transformer, Params=86M, FLOPs=55.4B, Train size=224, Test size=3842021.06 | 83.6 | |
| ResNet-34Training procedure=A12021.10 | 83.4 | |
| ViT-L/16 [14]Architecture=Transformer, Params=307M, FLOPs=191B, Train size=224, Test size=3842021.06 | 82.2 | |
| ViT-TiTraining procedure=A12021.10 | 82.1 | |
| ResNet-18Training procedure=A12021.10 | 79.4 |