Skip to first unread message
Julian Kolarz
Feb 1, 2017, 7:06:23 AM2/1/17
to Caffe Users
Hi all,
I have a question regarding the loss layer and how the loss is calculated. So i have a network that output a 1x18x18 matrix (blob). My groundtruth (label) is also a 1x18x18 matrix. I am using the Euclidean Loss Layer for training. Now my question. How is the loss calculated ?
Example with 1x3x3 matrix:
Label: Output (from net)
0 0 1 0 1 0
0 1 0 0 1 0
0 0 1 0 0 1
Now is the loss in the example 2 ? Or is it 0, because sum(label) = 3 and sum(output) = 3 ?
I would appreciated it, if somebody could give me a short answer to this :-)
I have a question regarding the loss layer and how the loss is calculated. So i have a network that output a 1x18x18 matrix (blob). My groundtruth (label) is also a 1x18x18 matrix. I am using the Euclidean Loss Layer for training. Now my question. How is the loss calculated ?
Example with 1x3x3 matrix:
Label: Output (from net)
0 0 1 0 1 0
0 1 0 0 1 0
0 0 1 0 0 1
Now is the loss in the example 2 ? Or is it 0, because sum(label) = 3 and sum(output) = 3 ?
I would appreciated it, if somebody could give me a short answer to this :-)
Reply all
Reply to author
Forward
0 new messages