search

kuangliu / pytorch-retinanet

RetinaNet in PyTorch
993 stars 249 forks source link

focal loss back propagation #12

Closed hengck23 closed 7 years ago

hengck23 commented 7 years ago

This is not an issue but a question.

I think the the term (1-p)^gamma and p^gamma in focal loss are for weighing only. They should not be back propagated during gradient descent. Am I correct?

If so, do you need to detach() your variables for computing the weight terms in your focal loss function?

hengck23 commented 7 years ago

after some checking, the weighing terms (1-p)^gamma and p^gamma are back propagated as well. you can refer to:

https://github.com/zimenglan-sysu-512/paper-note/blob/master/focal_loss.pdf https://github.com/unsky/focal-loss