Multi-label Image Classification on MS-COCO (val)
76.8F1 (C)SSGRL
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SSGRLLabel selection setting=All2019.08 | 76.8 | 83.8 | 89.9 | 68.5 | 79.7 | 91.3 | 70.8 | — | |
| SSGRLLabel selection setting=Top 32019.08 | 72.7 | 83.8 | 91.9 | 62.5 | 76.2 | 93.8 | 64.1 | — | |
| ResNet-SRNEvaluation Setting=All2017.02 | 71.2 | 77.1 | 81.6 | 65.4 | 75.8 | 82.7 | — | — | |
| ResNet-SRNLabel selection setting=All2019.08 | 71.2 | 77.1 | 81.6 | 65.4 | 75.8 | 82.7 | 69.9 | — | |
| ResNet-SRN-attEvaluation Setting=All2017.02 | 70 | 76.1 | 81.2 | 63.3 | 75 | 84.1 | — | — | |
| ResNet-SRN-attLabel selection setting=All2019.08 | 70 | 76.1 | 81.2 | 63.3 | 75 | 84.1 | 67.7 | — | |
| ResNet-101-semanticEvaluation Setting=All2017.02 | 69.9 | 75.5 | 81.1 | 63.8 | 74.8 | 82.1 | — | — | |
| ResNet-107Evaluation Setting=All2017.02 | 69.7 | 75.4 | 80.9 | 63.7 | 74.5 | 82.1 | — | — | |
| ResNet-101Evaluation Setting=All2017.02 | 69.5 | 75.2 | 80.8 | 63.4 | 74.4 | 82.2 | — | — | |
| KD-WSDLabel selection setting=All2019.08 | 69.2 | 74.6 | — | — | 74 | — | — | — | |
| ResNet-SRNEvaluation Setting=top-32017.02 | 67.4 | — | 85.2 | 58.8 | 72.9 | 87.4 | — | — | |
| RDARLabel selection setting=Top 32019.08 | 67.4 | 73.4 | 79.1 | 58.7 | 72 | 84 | 63 | — | |
| ResNet-SRNLabel selection setting=Top 32019.08 | 67.4 | 77.1 | 85.2 | 58.8 | 72.9 | 87.4 | 62.5 | — | |
| KD-WSDLabel selection setting=Top 32019.08 | 66.8 | 74.6 | — | — | 72.7 | — | — | — | |
| ResNet-SRN-attEvaluation Setting=top-32017.02 | 66.3 | — | 85.8 | 57.5 | 72.1 | 88.1 | — | — | |
| ResNet-SRN-attLabel selection setting=Top 32019.08 | 66.3 | 76.1 | 85.8 | 57.5 | 72.1 | 88.1 | 61.1 | — | |
| ResNet-101-semanticEvaluation Setting=top-32017.02 | 66.2 | — | 84.3 | 57.7 | 72 | 86.3 | — | — | |
| RARLLabel selection setting=Top 32019.08 | 66.2 | — | 78.8 | 57.2 | 71.1 | 84 | 61.6 | — | |
| ResNet-107Evaluation Setting=top-32017.02 | 66.1 | — | 84.4 | 57.6 | 71.8 | 86.4 | — | — | |
| ResNet-101Evaluation Setting=top-32017.02 | 65.9 | — | 84.3 | 57.4 | 71.7 | 86.5 | — | — | |
| RLSDLabel selection setting=Top 32019.08 | 62 | — | 67.6 | 57.2 | 66.5 | 70.1 | 63.4 | — | |
| CNN-RNNEvaluation Setting=top-32017.02 | 60.4 | — | 66 | 55.6 | 67.8 | 69.2 | — | — | |
| CNN-RNNLabel selection setting=Top 32019.08 | 60.4 | — | 66 | 55.6 | 67.8 | 69.2 | 66.4 | — | |
| WARPEvaluation Setting=top-32017.02 | 55.7 | — | 59.3 | 52.5 | 60.7 | 59.8 | — | — | |
| WARPLabel selection setting=Top 32019.08 | 55.7 | — | 59.3 | 52.5 | 60.7 | 59.8 | 61.4 | — | |
| Mixer-B-16Architecture=Mixer-B-16, Pretraining=Official ImageNet-21K2021.04 | — | — | — | — | — | — | — | 74.1 | |
| Mixer-B-16 (Semantic Softmax)Architecture=Mixer-B-16, Pretraining=Our ImageNet-21K (Semantic Softmax)2021.04 | — | — | — | — | — | — | — | 80.9 | |
| ViT-B-16Architecture=ViT-B-16, Pretraining=Official ImageNet-21K2021.04 | — | — | — | — | — | — | — | 81.1 | |
| ViT-B-16 (Semantic Softmax)Architecture=ViT-B-16, Pretraining=Our ImageNet-21K (Semantic Softmax)2021.04 | — | — | — | — | — | — | — | 82.6 |