Mask_RCNN
Mask_RCNN copied to clipboard
rpn_bbox_loss 0.4+ and others all less than 0.05
When I train the network, rpn_bbox_loss is too high, up to 0.4+. and others all ok less than 0.05 , my env is tf2.4.1 + python3.6, Anyone has the idea?
have you tried playing with the anchor size and ratio? I ran into the same issue in previous cases. Some can be solved and the rpn_bbox loss can be further lowered. But others won't go lower.
Here is the same problem i am dealing with
Here is the same problem i am dealing with
Have you been able to solve this problem? I'm dealing with this problem too but the loss is higher in my case. I could really use the advice. Thanks
Hello , I am having the same problem. I had change the Anchors size and also learning rate but rpn_bbox_loss stays high. does anyone have an insight how we can solve this issue ? or some hints? I would be gratful, thanks.