Why BCELoss for xy?
ccqu opened this issue · comments
Chengchao Qu commented
Hi,
I think the original loss for xy is L2, so why in your code is BCELoss used?
https://github.com/DeNA/PyTorch_YOLOv3/blob/master/models/yolo_layer.py#L192
Thanks in advance!
Lukas Struppek commented
Hi,
I am having the same question. The paper just mentioned BCE Loss for class prediction.
Yihui "Ray" Ren commented
add me to the list :D
same question here.