snapshot...

modules/deeplearning/srcnn_l1b_l2.py

@@ -66,14 +66,8 @@ y_134 = np.arange(134)
 x_64 = np.arange(64)
 y_64 = np.arange(64)
 
-x_134_2 = slice(3, 131, 2)
-y_134_2 = slice(3, 131, 2)
-# x_134_2 = x_134[3:131:2]
-# y_134_2 = y_134[3:131:2]
-
-t = np.arange(0, 64, 0.5)
-s = np.arange(0, 64, 0.5)
-
+# x_134_2 = slice(3, 131, 2)
+# y_134_2 = slice(3, 131, 2)
 x_128_2 = slice(3, 131, 2)
 y_128_2 = slice(3, 131, 2)
 x_128 = slice(3, 131)
@@ -85,17 +79,22 @@ y_128 = slice(3, 131)
 # y_128 = y_134[3:131]
 
 
-#----------- New
-# x_134_2 = x_134[1:134:2]
-# y_134_2 = y_134[1:134:2]
-# t = np.arange(1, 66, 0.5)
-# s = np.arange(1, 66, 0.5)
-#--------------------------
+x_134_2 = slice(1, 134, 2)
+y_134_2 = slice(1, 134, 2)
 
-slc_x_2 = x_128_2
-slc_y_2 = y_128_2
+# slc_x_2 = x_128_2
+# slc_y_2 = y_128_2
+# slc_x = x_128
+# slc_y = y_128
+# t = np.arange(0, 64, 0.5)
+# s = np.arange(0, 64, 0.5)
+
+slc_x_2 = x_134_2
+slc_y_2 = y_134_2
 slc_x = x_128
 slc_y = y_128
+t = np.arange(1, 66, 0.5)
+s = np.arange(1, 66, 0.5)
 
 
 def build_residual_conv2d_block(conv, num_filters, block_name, activation=tf.nn.relu, padding='SAME',
@@ -406,9 +405,10 @@ class SRCNN:
 
         input_2d = self.inputs[0]
         print('input: ', input_2d.shape)
-        # conv = tf.keras.layers.Conv2D(num_filters, kernel_size=5, strides=1, padding='VALID', activation=None)(input_2d)
-        conv = input_2d
-        print('input: ', conv.shape)
+        input_2d = tf.keras.layers.Conv2D(num_filters, kernel_size=3, strides=1, padding='VALID', activation=None)(input_2d)
+        # conv = input_2d
+        # print('input: ', conv.shape)
+        print('input: ', input_2d.shape)
 
         conv = conv_b = tf.keras.layers.Conv2D(num_filters, kernel_size=3, kernel_initializer='he_uniform', activation=activation, padding='SAME')(input_2d)
         print(conv.shape)