fix: remove unnecessary register_hook() call & pre-commit

ppcls/arch/backbone/legendary_models/esnet.py

@@ -217,7 +217,8 @@ class ESNet(TheseusLayer):
                  class_num=1000,
                  scale=1.0,
                  dropout_prob=0.2,
-                 class_expand=1280):
+                 class_expand=1280,
+                 return_patterns=None):
         super().__init__()
         self.scale = scale
         self.class_num = class_num
@@ -268,6 +269,9 @@ class ESNet(TheseusLayer):
         self.flatten = nn.Flatten(start_axis=1, stop_axis=-1)
         self.fc = Linear(self.class_expand, self.class_num)
 
+        if return_patterns is not None:
+            self.update_res(return_patterns)
+
     def forward(self, x):
         x = self.conv1(x)
         x = self.max_pool(x)

ppcls/arch/backbone/legendary_models/hrnet.py

@@ -244,7 +244,7 @@ class HighResolutionModule(TheseusLayer):
 
         for i in range(len(num_filters)):
             self.basic_block_list.append(
-                nn.Sequential(*[
+                nn.Sequential(* [
                     BasicBlock(
                         num_channels=num_filters[i],
                         num_filters=num_filters[i],
@@ -367,7 +367,11 @@ class HRNet(TheseusLayer):
         model: nn.Layer. Specific HRNet model depends on args.
     """
 
-    def __init__(self, width=18, has_se=False, class_num=1000, return_patterns=None):
+    def __init__(self,
+                 width=18,
+                 has_se=False,
+                 class_num=1000,
+                 return_patterns=None):
         super().__init__()
 
         self.width = width
@@ -394,7 +398,7 @@ class HRNet(TheseusLayer):
             stride=2,
             act="relu")
 
-        self.layer1 = nn.Sequential(*[
+        self.layer1 = nn.Sequential(* [
             BottleneckBlock(
                 num_channels=64 if i == 0 else 256,
                 num_filters=64,
@@ -458,7 +462,6 @@ class HRNet(TheseusLayer):
             weight_attr=ParamAttr(initializer=Uniform(-stdv, stdv)))
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, x):
         x = self.conv_layer1_1(x)

ppcls/arch/backbone/legendary_models/inception_v3.py

@@ -498,7 +498,6 @@ class Inception_V3(TheseusLayer):
             bias_attr=ParamAttr())
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, x):
         x = self.inception_stem(x)

ppcls/arch/backbone/legendary_models/mobilenet_v1.py

@@ -128,7 +128,7 @@ class MobileNet(TheseusLayer):
                     [int(512 * scale), 512, 1024, 512, 2],
                     [int(1024 * scale), 1024, 1024, 1024, 1]]
 
-        self.blocks = nn.Sequential(*[
+        self.blocks = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=params[0],
                 num_filters1=params[1],
@@ -147,7 +147,6 @@ class MobileNet(TheseusLayer):
             weight_attr=ParamAttr(initializer=KaimingNormal()))
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, x):
         x = self.conv(x)

ppcls/arch/backbone/legendary_models/mobilenet_v3.py

@@ -202,7 +202,6 @@ class MobileNetV3(TheseusLayer):
         self.fc = Linear(self.class_expand, class_num)
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, x):
         x = self.conv(x)

ppcls/arch/backbone/legendary_models/pp_lcnet.py

@@ -171,7 +171,8 @@ class PPLCNet(TheseusLayer):
                  scale=1.0,
                  class_num=1000,
                  dropout_prob=0.2,
-                 class_expand=1280):
+                 class_expand=1280,
+                 return_patterns=None):
         super().__init__()
         self.scale = scale
         self.class_expand = class_expand
@@ -182,7 +183,7 @@ class PPLCNet(TheseusLayer):
             num_filters=make_divisible(16 * scale),
             stride=2)
 
-        self.blocks2 = nn.Sequential(*[
+        self.blocks2 = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=make_divisible(in_c * scale),
                 num_filters=make_divisible(out_c * scale),
@@ -192,7 +193,7 @@ class PPLCNet(TheseusLayer):
             for i, (k, in_c, out_c, s, se) in enumerate(NET_CONFIG["blocks2"])
         ])
 
-        self.blocks3 = nn.Sequential(*[
+        self.blocks3 = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=make_divisible(in_c * scale),
                 num_filters=make_divisible(out_c * scale),
@@ -202,7 +203,7 @@ class PPLCNet(TheseusLayer):
             for i, (k, in_c, out_c, s, se) in enumerate(NET_CONFIG["blocks3"])
         ])
 
-        self.blocks4 = nn.Sequential(*[
+        self.blocks4 = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=make_divisible(in_c * scale),
                 num_filters=make_divisible(out_c * scale),
@@ -212,7 +213,7 @@ class PPLCNet(TheseusLayer):
             for i, (k, in_c, out_c, s, se) in enumerate(NET_CONFIG["blocks4"])
         ])
 
-        self.blocks5 = nn.Sequential(*[
+        self.blocks5 = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=make_divisible(in_c * scale),
                 num_filters=make_divisible(out_c * scale),
@@ -222,7 +223,7 @@ class PPLCNet(TheseusLayer):
             for i, (k, in_c, out_c, s, se) in enumerate(NET_CONFIG["blocks5"])
         ])
 
-        self.blocks6 = nn.Sequential(*[
+        self.blocks6 = nn.Sequential(* [
             DepthwiseSeparable(
                 num_channels=make_divisible(in_c * scale),
                 num_filters=make_divisible(out_c * scale),
@@ -248,6 +249,9 @@ class PPLCNet(TheseusLayer):
 
         self.fc = Linear(self.class_expand, class_num)
 
+        if return_patterns is not None:
+            self.update_res(return_patterns)
+
     def forward(self, x):
         x = self.conv1(x)
 

ppcls/arch/backbone/legendary_models/resnet.py

@@ -340,7 +340,6 @@ class ResNet(TheseusLayer):
         self.data_format = data_format
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, x):
         with paddle.static.amp.fp16_guard():

ppcls/arch/backbone/legendary_models/vgg.py

@@ -111,7 +111,11 @@ class VGGNet(TheseusLayer):
         model: nn.Layer. Specific VGG model depends on args.
     """
 
-    def __init__(self, config, stop_grad_layers=0, class_num=1000, return_patterns=None):
+    def __init__(self,
+                 config,
+                 stop_grad_layers=0,
+                 class_num=1000,
+                 return_patterns=None):
         super().__init__()
 
         self.stop_grad_layers = stop_grad_layers
@@ -139,7 +143,6 @@ class VGGNet(TheseusLayer):
         self.fc3 = Linear(4096, class_num)
         if return_patterns is not None:
             self.update_res(return_patterns)
-            self.register_forward_post_hook(self._return_dict_hook)
 
     def forward(self, inputs):
         x = self.conv_block_1(inputs)