TEM loss is not really decreasing

[yt2639]

Hey @JJBOY, thanks for your implementation and I found it extremely useful. I now have a problem during training, the PEM loss is decreasing normally, but TEM loss is not really decreasing, even it's increasing. Do you happen to have experience on how this may happen?

Thanks!

[JJBOY]

Thanks for your interests. TEM and PEM loss can not really reflect the final performance (e.g., AUC). Therefore, they are not of great reference value. They are printed just to see if training is going on properly.

------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2021年7月29日(星期四) 晚上9:00 收件人: @.>; 抄送: "Hello @.>; @.>; 主题: [JJBOY/BMN-Boundary-Matching-Network] TEM loss is not really decreasing (#44)

Hey @JJBOY, thanks for your implementation and I found it extremely useful. I now have a problem during training, the PEM loss is decreasing normally, but TEM loss is not really decreasing, even it's increasing. Do you happen to have experience on how this may happen?

Thanks!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[jojolee123]

i met this question too,loss didn't decrease on my own dataset,and even the official dataset.both of them the train loss is decreasing but val loss is not,how should i deal with it

[jojolee123]

so why the loss is not great reference value,and how cani make sure that my training is going on properly??some visualization?or some internal values?? THX