Object Detection on Objects365 v2 (val)
38.6AP50Supervised
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| SupervisedEvaluation Protocol=Supervised, Prompt Type=N/A2023.02 | 38.6 | 25.6 | — | 28 | 16 | 28.1 | 36.7 | |
| SupervisedZero-shot transfer=false, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 38.6 | 25.6 | — | 28 | 16 | 28.1 | 36.7 | |
| CoT-PLZero-shot transfer=true, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 22.7 | 15.1 | — | 15.9 | 6.1 | 14.4 | 22.5 | |
| Ourstraining_dataset=LVIS, base_model=Detic-base2022.07 | 22.3 | — | — | — | — | — | — | |
| LBPZero-shot transfer=true, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 21.8 | 14.3 | — | 15.1 | 5.5 | 13.7 | 21.6 | |
| Detictraining_dataset=LVIS, implementation=official2022.07 | 21.7 | — | — | — | — | — | — | |
| BARONEvaluation Protocol=Open-Vocabulary, Prompt Type=Learned2023.02 | 21 | 13.6 | — | 14.5 | 5 | 13.1 | 20.7 | |
| BARONZero-shot transfer=true, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 21 | 13.6 | — | 14.5 | 5 | 13.1 | 20.7 | |
| BARONEvaluation Protocol=Open-Vocabulary, Prompt Type=Hand-crafted2023.02 | 20 | 13.2 | — | 14 | 4.8 | 12.7 | 20.1 | |
| Detic-basetraining_dataset=LVIS, implementation=official2022.07 | 19.2 | — | — | — | — | — | — | |
| DetProEvaluation Protocol=Open-Vocabulary, Prompt Type=Learned2023.02 | 18.8 | 12.1 | — | 12.9 | 4.5 | 11.5 | 18.6 | |
| DetProZero-shot transfer=true, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 18.8 | 12.1 | — | 12.9 | 4.5 | 11.5 | 18.6 | |
| ViLDtraining_dataset=LVIS2022.07 | 18.2 | — | — | — | — | — | — | |
| ViLD*Evaluation Protocol=Open-Vocabulary, Prompt Type=Hand-crafted2023.02 | 17.8 | 11.5 | — | 12.3 | 4.2 | 11.1 | 17.8 | |
| ViLDZero-shot transfer=true, Trained on=OV-LVIS, Fine-tuning=false2025.10 | 17.8 | 11.5 | — | 12.3 | 4.2 | 11.1 | 17.8 | |
| ViLD-texttraining_dataset=LVIS2022.07 | 11.1 | — | — | — | — | — | — | |
| Box-SupervisedTraining dataset=LVIS-all, Supervision source=Box-supervised, Evaluation protocol=Zero-shot2022.01 | — | 19.1 | 14 | — | — | — | — | |
| Dataset-specific oraclesSupervision source=Dataset-specific box labels, Evaluation protocol=Supervised2022.01 | — | 31.2 | 22.5 | — | — | — | — | |
| DeticBackbone=Swin-B, Supervision source=IN-L (ImageNet-L) image-level, Evaluation protocol=Zero-shot, Loss function=Modified Federated Loss2022.01 | — | 21.2 | 17.8 | — | — | — | — | |
| DeticBackbone=Swin-B, Supervision source=IN-21k (ImageNet-21k) image-level, Evaluation protocol=Zero-shot, Loss function=Modified Federated Loss2022.01 | — | 21.5 | 20 | — | — | — | — |