lee-zq/3DUNet-Pytorch

TrainLoss decrease but ValLoss doesn't

Opened this issue · 2 comments

TrainLoss decrease but ValLoss doesn't

I use this code in my own wholeheart substructure segmentation,but the train loss decrease faster while valloss doesn't,how to solve this problem?

I use this code in my own wholeheart substructure segmentation,but the train loss decrease faster while valloss doesn't,how to solve this problem?

hello, I encountered the same problem, have you solved it?