hehaodele / cida Goto Github PK

It seems that the best_acc_target was never updated? And only the last model was saved. I was wondering whether we should update the best_acc_target and use the best model to get the performance?

我发现代码里面有一个问题，你没有划分训练集和测试集？你好像用的是同一个。

Could you please release the code for medical datasets? Thanks!

Hi Hao, I watched your presentation but missed some points, could I ask what the ratio of labeled source and unlabeled target in your training, and for unlabeled target prediction \hat{y}, how do you regress it? Thanks.