Closed FXLYZ closed 1 year ago
Hi,
sry for the late reply. I we used the standard evaluation metrics of OpenImages VRD Challenge. You can find it in the RelDN repo: https://github.com/NVIDIA/ContrastiveLosses4VRD/blob/master/lib/datasets_rel/ap_eval_rel.py
Hi, Can you share the code of computing the average precision of predicates? Thank you! 😊