per-pixel multinomial logistic loss

123 views

Skip to first unread message

Tsatsin Petr

unread,

Oct 18, 2016, 1:25:08 AM10/18/16

to torch7

I was wondering if there is a Criterion that computes the segmentation loss similar to the one used in Fully Convolutional Networks for Semantic Segmentation (arXiv:1411.4038v2)? Or are there any examples how it can be implemented?

Kiran Vaidhya

unread,

Oct 22, 2016, 10:03:46 AM10/22/16

to torch7

Try nn.SpatialClassNLLCriterion() or cudnn.SpatialCrossEntropyCriterion()

Reply all

Reply to author

Forward

0 new messages