Fu_Yimin comments

Repositories
Issues
Comments

Results 3 comments of


                                            Fu_Yimin

trafficstars

请问在up loss中关于background类别的处理。

> 您好，我想问一下这里的self.num_class是know_class+unknow_class+bg嘛我看到代码分类训练是是self.cls_score = nn.Linear( self.cls_score.in_features, self.num_classes + 1, bias=False)不太明白未知类的概率是怎么得到的。感谢同问

About loading the ImageNet Pretrain

Thanks for your reply

question about the setting of lambd(constant) for grad_reverse

> It is a common operation when using gradient reversal. See Eq. 14 in the following paper. https://arxiv.org/pdf/1409.7495.pdf thanks for your reply.