Object Detection on COCO-standard 0.5% labeled 2017 (train)
21.26mAPListen2Student
Evaluation results
| Method | Links | |
|---|---|---|
| Listen2StudentBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=32/322022.06 | 21.26 | |
| Listen2StudentBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=8/402022.06 | 21.02 | |
| Virtual Category Learning (with tricks)Configuration=Training tricks in Soft Teacher2022.07 | 19.46 | |
| Virtual Category LearningConfiguration=Same settings as Unbiased Teacher2022.07 | 18.12 | |
| Listen2StudentBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=8/82022.06 | 17.51 | |
| Unbiased TeacherBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=32/322022.06 | 16.94 | |
| Unbiased Teacher2022.07 | 16.94 | |
| Soft TeacherOfficial code=true2022.07 | 15.04 | |
| Unbiased TeacherBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=8/82022.06 | 14.36 | |
| STACBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=8/82022.06 | 9.78 | |
| STAC2022.07 | 9.78 | |
| CSDBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based, Batch Size (labeled/unlabeled)=8/82022.06 | 7.41 | |
| CSD2022.07 | 7.41 | |
| SupervisedBackbone=FasterRCNN-ResNet50, Detector Type=Anchor-based2022.06 | 6.83 | |
| Supervised2022.07 | 6.83 |