how does the loss converge

happynear / AMSoftmax

A simple yet effective loss function for face verification.

MIT License

487 stars 130 forks source link

how does the loss converge #25

Closed jyshee closed 5 years ago

jyshee commented 6 years ago

Does the AMS loss have the similar converge curve with softmax loss? In my exps, the AMS loss (set m=1) changes little during training, even after lots of iters.

happynear commented 6 years ago

Loss is meaningless in AMS. Maybe you can output the accuracy, target logit etc to observe the convergence of the network.