Hi.
I want to implement early stopping for faster rcnn
and mask rcnn
. But faster rcnn
do not show validation loss
during training and just print validation done!
also mask rcnn
do not print validation loss
.
During the faster_rcnn training, in each epoch, there are 4 kinds of loss printing.
You can leverage them for early stopping.
For specific epoch, mAPs are computed. It will also show AP, precision, recall and RPN_recall.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.