search

CuriousAI / mean-teacher

A state-of-the-art semi-supervised method for image recognition
https://arxiv.org/abs/1703.01780
Other
1.58k stars 331 forks source link

About the loss compute #39

Open Howardqlz opened 4 years ago

Howardqlz commented 4 years ago

Hi, i am wonder about this loss: class_loss = class_criterion(class_logit, target_var) / minibatch_size since this loss ignore some samples(no_lable), why here still use the minibatch_size not the labeled_size?