yolo2-pytorch icon indicating copy to clipboard operation
yolo2-pytorch copied to clipboard

Published 20 hours ago • verified publisher icon longcw

About train loss

Open zhengnianqing opened this issue 7 years ago • 0 comments

I have a question about the code in the function of def _process_batch in the darknet.py : the _boxes[:,:,0:2] = 0.5 _boxes[:,:,2:4] =1.0 .Thus the noobject will have the box regression loss. Why does noobject cause loss?

zhengnianqing avatar May 22 '18 11:05 zhengnianqing