コード例 #1
0
 def forward(self, x, with_feature=True):
     f_list, x = unpack_feature(x)
     if self.use_res_connect:
         ret = x + self.conv(x)
     else:
         ret = self.conv(x)
     return pack_feature(f_list, ret, with_feature)
コード例 #2
0
 def forward(self, x, with_feature=False):
     x = self.features(x)
     f_list, x = unpack_feature(x)
     x = x.mean([2, 3])
     f_list.append(x)
     x = self.classifier(x)
     return pack_feature(f_list, x, with_feature)
コード例 #3
0
 def forward(self, x, with_feature=False):
     features = self.features(x)
     f_list, features = unpack_feature(features)
     out = F.relu(features, inplace=True)
     out = F.adaptive_avg_pool2d(out, (1, 1)).view(features.size(0), -1)
     f_list.append(out)
     out = self.classifier(out)
     return pack_feature(f_list, out, with_feature)
コード例 #4
0
    def forward(self, x, with_feature=True):
        f_list, x_last = unpack_feature(x)

        x = F.relu(self.bn1(self.conv1(x_last)))
        x = self.bn2(self.conv2(x))
        x += self.shortcut(x_last)
        x = F.relu(x)

        return pack_feature(f_list, x, with_feature)
コード例 #5
0
ファイル: wrn.py プロジェクト: wukailu/DomainGenralization
 def forward(self, x, with_feature=True):
     f_list, x = unpack_feature(x)
     if not self.equalInOut:
         x = self.relu1(self.bn1(x))
     else:
         out = self.relu1(self.bn1(x))
     out = self.relu2(self.bn2(self.conv1(out if self.equalInOut else x)))
     if self.droprate > 0:
         out = F.dropout(out, p=self.droprate, training=self.training)
     out = self.conv2(out)
     out = torch.add(x if self.equalInOut else self.convShortcut(x), out)
     return pack_feature(f_list, out, with_feature)
コード例 #6
0
ファイル: wrn.py プロジェクト: wukailu/DomainGenralization
    def forward(self, x, with_feature=False):
        out = self.conv1(x)
        out = self.block1(out)
        out = self.block2(out)
        out = self.block3(out)

        f_list, out = unpack_feature(out)
        out = self.relu(self.bn1(out))
        out = F.avg_pool2d(out, 8)
        out = out.view(-1, self.nChannels)

        f_list.append(out)

        out = self.fc(out)

        return pack_feature(f_list, out, with_feature)
コード例 #7
0
ファイル: resnet.py プロジェクト: wukailu/DomainGenralization
    def forward(self, x, with_feature=True):
        f_list, x = unpack_feature(x)

        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)

        if self.downsample is not None:
            identity = self.downsample(x)

        out += identity
        out = self.relu(out)

        return pack_feature(f_list, out, with_feature)
コード例 #8
0
    def forward(self, x, with_feature=False):
        x = F.relu(self.bn1(self.conv1(x)))

        # x = pack_feature(*unpack_feature(x))

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)

        f_list, x = unpack_feature(x)

        x = F.avg_pool2d(x, x.size()[3])
        x = x.view(x.size(0), -1)

        f_list.append(x)

        x = self.linear(x)

        return pack_feature(f_list, x, with_feature)
コード例 #9
0
ファイル: resnet.py プロジェクト: wukailu/DomainGenralization
    def forward(self, x, with_feature=False):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        f_list, x = unpack_feature(x)

        x = self.avgpool(x)
        x = x.reshape(x.size(0), -1)

        f_list.append(x)

        x = self.fc(x)

        return pack_feature(f_list, x, with_feature)