OV-DINO: Unified Open-Vocabulary Detection with Language-Aware Selective Fusion

Benchmark Model Rank Results
zero-shot-object-detection-on-lvis-v1-0-minivalOV-DINO-T (without LVIS data, swin tiny)#5AP: 40.1
zero-shot-object-detection-on-lvis-v1-0-valOV-DINO-T (without LVIS data, swin tiny)#5AP: 32.9
zero-shot-object-detection-on-mscocoOV-DINO-T (without COCO data)#4AP: 50.6