Object Detection on COCO Open Vocabulary
48.3AP NovelOV-DQUO+DeCLIP
Evaluation results
| Method | Links | |
|---|---|---|
| OV-DQUO+DeCLIPSupervision=CLIP, Backbone=ViT-L/142025.05 | 48.3 | |
| F-ViT+DeCLIPSupervision=CLIP, Backbone=ViT-L/142025.05 | 46.2 | |
| OV-DQUO+DeCLIPSupervision=CLIP, Backbone=ViT-B/162025.05 | 46.1 | |
| OV-DQUOSupervision=CLIP, Backbone=RN50x42025.05 | 45.6 | |
| CLIPSelfSupervision=CLIP, Backbone=ViT-L/142025.05 | 44.3 | |
| CORA+Supervision=CLIP, Backbone=RN50x42025.05 | 41.7 | |
| F-ViT+DeCLIPSupervision=CLIP, Backbone=ViT-B/162025.05 | 41.1 | |
| RegionCLIPSupervision=Captions, Backbone=RN50x42025.05 | 39.3 | |
| OV-DQUOSupervision=CLIP, Backbone=RN502025.05 | 39.2 | |
| CLIPSelfSupervision=CLIP, Backbone=ViT-B/162025.05 | 37.6 | |
| SAS-DetSupervision=CLIP, Backbone=RN502025.05 | 37.4 | |
| CFM-VITSupervision=CLIP, Backbone=ViT-L/162025.05 | 34.1 | |
| BARON-KDSupervision=CLIP, Backbone=RN502025.05 | 34 | |
| RO-VITSupervision=CLIP, Backbone=ViT-L/162025.05 | 33 | |
| OV-DETRSupervision=CLIP, Backbone=RN502025.05 | 29.4 | |
| DeticSupervision=Caption, Backbone=RN502025.05 | 27.8 | |
| ViLDSupervision=CLIP, Backbone=RN502025.05 | 27.6 |