To miss-attend is to misalign! Residual Self-Attentive Feature Alignment for Adapting Object Detectors

Benchmark Model Rank Results
unsupervised-domain-adaptation-on-bdd100k-toILLUME#3mAP: 29.6
unsupervised-domain-adaptation-on-cityscapes-1ILLUME#7mAP@0.5: 43.8
unsupervised-domain-adaptation-on-sim10k-to-3ILLUME#11mAP@0.5: 53.1