Object Detection on LVIS full (val)
50.4AP all (old)OWL-ST+FT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| OWL-ST+FTBackbone=CLIP L/14, Vocabulary Type=Human-curated vocabulary2023.06 | 50.4 | 52 | 45.9 | 48.5 | |
| OWL-ST+FTBackbone=CLIP L/14, Vocabulary Type=Open vocabulary2023.06 | 49.4 | 51.1 | 44.6 | 47.4 | |
| OWL-ST+FTBackbone=SigLIP G/14, Vocabulary Type=Open vocabulary2023.06 | 47 | — | 47.2 | — | |
| OWL-ST+FTBackbone=CLIP B/16, Vocabulary Type=Human-curated vocabulary2023.06 | 45.6 | 46.7 | 40.5 | 42.5 | |
| 3WaysBackbone=NFNet-F6, Vocabulary Type=Open vocabulary2023.06 | 44.6 | — | 30.1 | — | |
| OWL-ST+FTBackbone=CLIP B/16, Vocabulary Type=Open vocabulary2023.06 | 41.8 | 43.2 | 36.2 | 39 | |
| 3WaysBackbone=NFNet-F0, Vocabulary Type=Open vocabulary2023.06 | 35.7 | — | 25.6 | — | |
| F-VLMBackbone=R50x64, Vocabulary Type=Open vocabulary2023.06 | 34.9 | — | 32.8 | — | |
| OWLBackbone=CLIP L/14, Vocabulary Type=Open vocabulary2023.06 | 34.6 | — | 31.2 | — | |
| OWL-STBackbone=SigLIP G/14, Vocabulary Type=Open vocabulary2023.06 | 33.7 | — | 37.5 | — | |
| OWL-STBackbone=CLIP L/14, Vocabulary Type=Open vocabulary2023.06 | 33.5 | 35.2 | 34.9 | 36.2 | |
| DeticBackbone=R50, Vocabulary Type=Human-curated vocabulary2023.06 | 32.4 | — | 24.6 | — | |
| RegionCLIPBackbone=R50x4, Vocabulary Type=Open vocabulary2023.06 | 32.3 | — | 22 | — | |
| F-VLMBackbone=R50x4, Vocabulary Type=Open vocabulary2023.06 | 28.5 | — | 26.3 | — | |
| OWLBackbone=CLIP B/16, Vocabulary Type=Open vocabulary2023.06 | 27.2 | — | 20.6 | — | |
| OWL-STBackbone=CLIP B/16, Vocabulary Type=Open vocabulary2023.06 | 27 | 28.6 | 29.6 | 30.3 | |
| DetCLIPv2Backbone=Swin-T, Vocabulary Type=Human-curated vocabulary2023.06 | — | 32.8 | — | 31 | |
| DetCLIPv2Backbone=Swin-L, Vocabulary Type=Human-curated vocabulary2023.06 | — | 36.6 | — | 33.3 |