easy-faster-rcnn.pytorch icon indicating copy to clipboard operation
easy-faster-rcnn.pytorch copied to clipboard

Published 20 hours ago • verified publisher icon potterhsu

L1 regression loss for proposals going to 0

Open peiherng97 opened this issue 6 years ago • 0 comments

Hi, I noticed that the smooth L1 regression loss for the fast r cnn goes to 0 when there are no proposals that have an iou of > 0.5. Is this deliberately done? The fg_indices are empty when there are no matches

https://github.com/potterhsu/easy-faster-rcnn.pytorch/blob/2c30c6d4ea57402c813294a499181b6ad710f858/model.py#L124-L129

peiherng97 avatar Jul 17 '19 16:07 peiherng97