[Question]About loss
alonelysnake opened this issue · comments
❓ Question
In my task, I found the train loss<=0 after several epochs. What caused this problem? Should I continue to train the model?
Yes, simply continue the training, it is normal that the loss becomes negative at some points. While most losses in nnDetection are positive, we simply use -GIoU
as the loss instead of 1-GIoU
to avoid unnecessary computations - the optimization target remains the same in both cases.
This issue is stale because it has been open for 30 days with no activity.
This issue was closed because it has been inactive for 14 days since being marked as stale.