revert

migrate.py

@@ -26,18 +26,18 @@ def migrate_model(new_model):
         new_layer = new_layers[i + 1]
         new_layer.set_weights(old_layer.get_weights())
 
-    flatten = old_model.get_layer('flatten')
-    f_dim = flatten.input_shape
-    print('f_dim: ' + str(f_dim))
-    old_dense1 = old_model.get_layer('dense1')
-    input_shape = old_dense1.input_shape
-    output_dim = old_dense1.get_weights()[1].shape[0]
-    print('output_dim: ' + str(output_dim))
-    W, b = old_dense1.get_weights()
-    shape = (7, 7, 512, output_dim)
-    new_W = W.reshape(shape)
-    new_conv6 = new_model.get_layer('conv6')
-    new_conv6.set_weights([new_W, b])
+    # flatten = old_model.get_layer('flatten')
+    # f_dim = flatten.input_shape
+    # print('f_dim: ' + str(f_dim))
+    # old_dense1 = old_model.get_layer('dense1')
+    # input_shape = old_dense1.input_shape
+    # output_dim = old_dense1.get_weights()[1].shape[0]
+    # print('output_dim: ' + str(output_dim))
+    # W, b = old_dense1.get_weights()
+    # shape = (7, 7, 512, output_dim)
+    # new_W = W.reshape(shape)
+    # new_conv6 = new_model.get_layer('conv6')
+    # new_conv6.set_weights([new_W, b])
 
     del old_model
 

model.py

@@ -51,9 +51,9 @@ def build_encoder_decoder():
     x = MaxPooling2D((2, 2), strides=(2, 2))(x)
 
     # Decoder
-    x = Conv2D(4096, (7, 7), activation='relu', padding='valid', name='conv6')(x)
-    x = BatchNormalization()(x)
-    x = UpSampling2D(size=(7, 7))(x)
+    # x = Conv2D(4096, (7, 7), activation='relu', padding='valid', name='conv6')(x)
+    # x = BatchNormalization()(x)
+    # x = UpSampling2D(size=(7, 7))(x)
 
     x = Conv2D(512, (1, 1), activation='relu', padding='same', name='deconv6', kernel_initializer='he_normal',
                bias_initializer='zeros')(x)