search

weiliu89 / caffe

Caffe: a fast open framework for deep learning.
http://caffe.berkeleyvision.org/
Other
4.77k stars 1.67k forks source link

prior_variance #629

Open achao2013 opened 7 years ago

achao2013 commented 7 years ago

hi, i want to ask why set prior_variance = [0.1, 0.1, 0.2, 0.2]. if i adjust it , what will happens to the result. Thanks a lot

weiliu89 commented 7 years ago

It gives the loc loss more weight.