Merge pull request WZMIAOMIAO#163 from WZMIAOMIAO/dev

update model

pytorch_classification/Test10_regnet/model.py

@@ -98,13 +98,13 @@ def __init__(self,
                               bias=False)
 
         self.bn = nn.BatchNorm2d(out_c)
-        self.act = act
+        self.act = act if act is not None else nn.Identity()
 
     def forward(self, x: Tensor) -> Tensor:
         x = self.conv(x)
         x = self.bn(x)
-        if self.act is not None:
-            x = self.act(x)
+        # if self.act is not None:
+        x = self.act(x)
         return x
 
 
@@ -120,15 +120,14 @@ def __init__(self,
         if drop_ratio > 0:
             self.dropout = nn.Dropout(p=drop_ratio)
         else:
-            self.dropout = None
+            self.dropout = nn.Identity()
 
         self.fc = nn.Linear(in_features=in_unit, out_features=out_unit)
 
     def forward(self, x: Tensor) -> Tensor:
         x = self.pool(x)
         x = torch.flatten(x, start_dim=1)
-        if self.dropout is not None:
-            x = self.dropout(x)
+        x = self.dropout(x)
         x = self.fc(x)
         return x
 
@@ -172,20 +171,20 @@ def __init__(self,
         if se_ratio > 0:
             self.se = SqueezeExcitation(in_c, out_c, se_ratio)
         else:
-            self.se = None
+            self.se = nn.Identity()
 
         self.conv3 = ConvBNAct(in_c=out_c, out_c=out_c, kernel_s=1, act=None)
         self.ac3 = nn.ReLU(inplace=True)
 
         if drop_ratio > 0:
             self.dropout = nn.Dropout(p=drop_ratio)
         else:
-            self.dropout = None
+            self.dropout = nn.Identity()
 
         if (in_c != out_c) or (stride != 1):
             self.downsample = ConvBNAct(in_c=in_c, out_c=out_c, kernel_s=1, stride=stride, act=None)
         else:
-            self.downsample = None
+            self.downsample = nn.Identity()
 
     def zero_init_last_bn(self):
         nn.init.zeros_(self.conv3.bn.weight)
@@ -195,15 +194,12 @@ def forward(self, x: Tensor) -> Tensor:
         x = self.conv1(x)
         x = self.conv2(x)
 
-        if self.se is not None:
-            x = self.se(x)
+        x = self.se(x)
         x = self.conv3(x)
 
-        if self.dropout is not None:
-            x = self.dropout(x)
+        x = self.dropout(x)
 
-        if self.downsample is not None:
-            shortcut = self.downsample(shortcut)
+        shortcut = self.downsample(shortcut)
 
         x += shortcut
         x = self.ac3(x)