search

seathiefwang / RankPose

RankPose: Learning Generalised Feature with Rank Supervision for Head Pose Estimation
29 stars 7 forks source link

It is seems something wrong in use nn.MarginRankingLoss #4

Open Etpoem opened 4 years ago

Etpoem commented 4 years ago

I The loss function of nn.MarginRankingLoss is "loss(x1,x2,y)=max(0,−y∗(x1−x2)+margin) " where y expect 1 or -1, but labels only has value in 0 or 1