wkentaro / pytorch-fcn

PyTorch Implementation of Fully Convolutional Networks. (Training code to reproduce the original result is available.)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Small question about score_pool3 and score_pool4 in FCN8

k4kuz0 opened this issue · comments

commented

Hi! I have a small question. Why is the score_pool4 and score_pool3 multiplied with 0.01 and 0.0001 respectively?

Can be seen here:
https://github.com/wkentaro/pytorch-fcn/blob/master/torchfcn/models/fcn8s.py#L222-L230

I don't understand the comment "# XXX: scaling to train at once"

commented

Never mind, I can see it is also done in the original implementation and thus this repo isn't the right place to ask this question.

Can be seen in original implementation here:
https://github.com/shelhamer/fcn.berkeleyvision.org/blob/master/voc-fcn8s-atonce/net.py#L65-L79