OV-DQUO: Open-Vocabulary DETR with Denoising Text Query Training and Open-World Unknown Objects Supervision

Open paper
Benchmark Model Rank Results
open-vocabulary-object-detection-on-lvis-v1-0OV-DQUO(ViT-L/14)#3AP novel-LVIS base training: 39.3
open-vocabulary-object-detection-on-lvis-v1-0OV-DQUO(ViT-B/16)#12AP novel-LVIS base training: 29.7
open-vocabulary-object-detection-on-mscocoOV-DQUO(RN50x4)#4AP 0.5: 45.6
open-vocabulary-object-detection-on-mscocoOV-DQUO(R50)#14AP 0.5: 39.2

Rank counts only results with a code link.