search

WeitaoVan / L-GM-loss

Implementation of our accepted CVPR 2018 paper "Rethinking Feature Distribution for Loss Functions in Image Classification"
MIT License
172 stars 34 forks source link

How to update the variance #3

Closed Merle314 closed 6 years ago

Merle314 commented 6 years ago

thank you very much for sharing your code. I want to know how you update the variance of each class, or just take a constant value?

WeitaoVan commented 6 years ago

For caffe, in the class_distance layer, set $update_sigma$ to true, and set lr_mult to a low ratio. Please check the instructions for this reposity.