Added a Conv1D layer at the end of both decoder and next_sequence_predictor

deepof/models.py

@@ -415,11 +415,6 @@ class SEQ_2_SEQ_GMVAE:
             Model_E4.append(BatchNormalization())
 
         # Decoder layers
-        Model_B1 = BatchNormalization()
-        Model_B2 = BatchNormalization()
-        Model_B3 = BatchNormalization()
-        Model_B4 = BatchNormalization()
-
         seq_D = [
             Dense(
                 self.DENSE_2,
@@ -506,6 +501,15 @@ class SEQ_2_SEQ_GMVAE:
             ),
             merge_mode=self.bidirectional_merge,
         )
+        Model_P4 = tf.keras.layers.Conv1D(
+            filters=self.CONV_filters,
+            kernel_size=5,
+            strides=1,
+            padding="same",
+            activation=self.dense_activation,
+            kernel_initializer=he_uniform(),
+            use_bias=True,
+        )
 
         # Phenotype classification layer
         Model_PC1 = Dense(
@@ -527,10 +531,6 @@ class SEQ_2_SEQ_GMVAE:
             Model_E2,
             Model_E3,
             Model_E4,
-            Model_B1,
-            Model_B2,
-            Model_B3,
-            Model_B4,
             Model_D1,
             Model_D2,
             Model_D3,
@@ -540,6 +540,7 @@ class SEQ_2_SEQ_GMVAE:
             Model_P1,
             Model_P2,
             Model_P3,
+            Model_P4,
             Model_PC1,
             Model_RC1,
         )
@@ -557,10 +558,6 @@ class SEQ_2_SEQ_GMVAE:
             Model_E2,
             Model_E3,
             Model_E4,
-            Model_B1,
-            Model_B2,
-            Model_B3,
-            Model_B4,
             Model_D1,
             Model_D2,
             Model_D3,
@@ -570,6 +567,7 @@ class SEQ_2_SEQ_GMVAE:
             Model_P1,
             Model_P2,
             Model_P3,
+            Model_P4,
             Model_PC1,
             Model_RC1,
         ) = self.get_layers(input_shape)
@@ -697,14 +695,14 @@ class SEQ_2_SEQ_GMVAE:
         g = Input(shape=self.ENCODING)
         generator = Sequential(Model_D1)(g)
         generator = Model_D2(generator)
-        generator = Model_B1(generator)
+        generator = BatchNormalization()(generator)
         generator = Model_D3(generator)
         generator = Model_D4(generator)
-        generator = Model_B2(generator)
+        generator = BatchNormalization()(generator)
         generator = Model_D5(generator)
-        generator = Model_B3(generator)
+        generator = BatchNormalization()(generator)
         generator = Model_D6(generator)
-        generator = Model_B4(generator)
+        generator = BatchNormalization()(generator)
         x_decoded_mean = Dense(
             tfpl.IndependentNormal.params_size(input_shape[2:]) // 2
         )(generator)
@@ -750,6 +748,7 @@ class SEQ_2_SEQ_GMVAE:
             predictor = BatchNormalization()(predictor)
             predictor = Model_P3(predictor)
             predictor = BatchNormalization()(predictor)
+            predictor = Model_P4(predictor)
             x_predicted_mean = Dense(
                 tfpl.IndependentNormal.params_size(input_shape[2:]) // 2
             )(predictor)