Why is local feature (lf) different in training phase and test phase?

[anonymoussss]

Hi~，
I notice that local feature (lf) variable is different in training phase and test phase. Specifically, in "AlignedReID/models/ResNet.py",

if not self.training:
            lf = self.horizon_pool(x)
        if self.aligned and self.training:
            lf = self.bn(x)
            lf = self.relu(lf)
            lf = self.horizon_pool(lf)
            lf = self.conv1(lf)
        if self.aligned or not self.training:
            lf = lf.view(lf.size()[0:3])
            lf = lf / torch.pow(lf,2).sum(dim=1, keepdim=True).clamp(min=1e-12).sqrt()

when training, 'lf' is 'self.horizon_pool+self.conv1', but during test, 'lf' is only 'self.horizon_pool'. Wouldn't this mismatch between the training phase and the testing phase make the results worse?