Object Detection on COCO-C
42mAPFAN-L-Hybrid
Evaluation results
| Method | Links | |||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| FAN-L-HybridEncoder Size=76.8M, Detector=Cascade Mask R-CNN, Pre-trained=ImageNet-22K2022.04 | 42 | 76.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FAN-L-HybridEncoder Size=76.8M, Detector=Cascade Mask R-CNN2022.04 | 40.6 | 75 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FAN-S-HybridEncoder Size=26.3M, Detector=Cascade Mask R-CNN2022.04 | 38.7 | 72.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FAN-S-HybridEncoder Size=26.3M, Detector=Mask R-CNN2022.04 | 35.5 | 72.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FAN-T-HybridEncoder Size=7.4M, Detector=Cascade Mask R-CNN2022.04 | 33.1 | 65.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| FAN-T-HybridEncoder Size=7.4M, Detector=Mask R-CNN2022.04 | 29.7 | 64.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Swin-TEncoder Size=28.0M, Detector=Mask R-CNN2022.04 | 29.3 | 63.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLOD-TTA*Detector=Grounding-DINO2025.10 | 27.3 | — | — | 21.1 | 22.2 | 21.2 | 21.4 | 12.9 | 19.7 | 9.9 | 32.3 | 36.9 | 47.8 | 47 | 31.6 | 30.1 | 25.9 | 29.2 | — | — | — | — | — | |
| DeiT-SEncoder Size=22.1M, Detector=Mask R-CNN2022.04 | 26.9 | 67.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| VLOD-TTADetector=Grounding-DINO2025.10 | 26.2 | — | — | 20.5 | 21.8 | 20.9 | 20.6 | 13.1 | 18.4 | 10.1 | 30.8 | 36.6 | 48.2 | 46.2 | 28.8 | 28.7 | 19.7 | 28.6 | — | — | — | — | — | |
| TTAOD-FDetector=Grounding-DINO2025.10 | 26 | — | — | 20.2 | 22 | 21.4 | 17.8 | 14.5 | 16.9 | 7.9 | 31.1 | 34.7 | 45.1 | 44.9 | 30.6 | 29.9 | 23.6 | 20.2 | — | — | — | — | — | |
| DATModel=YOLOv3-416 [66], Training Strategy=DAT (Ours)2022.09 | 23.38 | — | 57.86 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResNet101Encoder Size=44.1M, Detector=Mask R-CNN2022.04 | 23.3 | 55.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Det-AdvpropModel=YOLOv3-416 [66], Training Strategy=Det-Advprop2022.09 | 22.55 | — | 55.42 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DATModel=EffDet-Lite1-384 [65], Training Strategy=DAT (Ours)2022.09 | 21.32 | — | 65.99 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ResNet-50Encoder Size=25.4M, Detector=Mask R-CNN2022.04 | 21.3 | 53.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NormalModel=YOLOv3-416 [66], Training Strategy=Normal2022.09 | 21.19 | — | 52.58 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ZSDetector=Grounding-DINO2025.10 | 20.6 | — | — | 13.7 | 16 | 15 | 16.8 | 7.5 | 13.6 | 6.7 | 27.5 | 32.5 | 44.2 | 44.1 | 21.9 | 22.5 | 5.3 | 21.1 | — | — | — | — | — | |
| DATModel=YOLOv3-320 [66], Training Strategy=DAT (Ours)2022.09 | 20.55 | — | 57.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Det-AdvpropModel=EffDet-Lite1-384 [65], Training Strategy=Det-Advprop2022.09 | 20.21 | — | 63.51 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NormalModel=EffDet-Lite1-384 [65], Training Strategy=Normal2022.09 | 19.43 | — | 61.68 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Det-AdvpropModel=YOLOv3-320 [66], Training Strategy=Det-Advprop2022.09 | 19.34 | — | 54.13 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NormalModel=YOLOv3-320 [66], Training Strategy=Normal2022.09 | 18.39 | — | 51.21 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DATModel=EffDet-Lite0-320 [65], Training Strategy=DAT (Ours)2022.09 | 17.89 | — | 65.48 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Det-AdvpropModel=EffDet-Lite0-320 [65], Training Strategy=Det-Advprop2022.09 | 16.38 | — | 62.19 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| NormalModel=EffDet-Lite0-320 [65], Training Strategy=Normal2022.09 | 16.11 | — | 61 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| TTAOD-FTotal Params (Mil)=629.9, Tuned Params (Mil)=0.08, Memory (GB)=11.37, Latency (ms/img)=701.92025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 21.2 | 14.3 | 37 | 31.7 | 26 | |
| VLOD-TTATotal Params (Mil)=173.9, Tuned Params (Mil)=0.89, Memory (GB)=3.76, Latency (ms/img)=531.62025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 21.1 | 15.6 | 38.5 | 30.4 | 26.2 | |
| VLOD-TTA*Total Params (Mil)=173.9, Tuned Params (Mil)=0.95, Memory (GB)=3.92, Latency (ms/img)=567.32025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 21.5 | 15.9 | 38.6 | 32.8 | 27.3 | |
| ZSTotal Params (Mil)=172.9, Tuned Params (Mil)=0.00, Memory (GB)=1.36, Latency (ms/img)=210.62025.10 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 14.9 | 11.2 | 34.7 | 23 | 20.6 |