Define input_shape via Input call in Sequential model

This is only for MLP for now.
feranick · Oct 8, 2024 · 7458220 · 7458220
1 parent 1483bf5
commit 7458220
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 3 deletions.
diff --git a/SpectraKeras/SpectraKeras_CNN.py b/SpectraKeras/SpectraKeras_CNN.py
@@ -3,7 +3,7 @@
 '''
 **********************************************
 * SpectraKeras_CNN Classifier and Regressor
-* v2024.10.07.2
+* v2024.10.08.1
 * Uses: TensorFlow
 * By: Nicola Ferralis <[email protected]>
 **********************************************

diff --git a/SpectraKeras/SpectraKeras_MLP.py b/SpectraKeras/SpectraKeras_MLP.py
@@ -3,7 +3,7 @@
 '''
 **********************************************
 * SpectraKeras_MLP Classifier and Regressor
-* v2024.10.07.3
+* v2024.10.08.1
 * Uses: TensorFlow
 * By: Nicola Ferralis <[email protected]>
 **********************************************
@@ -304,10 +304,11 @@ def get_model():
                 amsgrad=False)
 
         model = keras.models.Sequential()
+        model.add(keras.Input(shape=(A.shape[1],)))
         for i in range(len(dP.HL)):
             model.add(keras.layers.Dense(dP.HL[i],
                 activation = 'relu',
-                input_dim=A.shape[1],
+                #input_dim=A.shape[1],
                 kernel_regularizer=keras.regularizers.l2(dP.l2)))
             model.add(keras.layers.Dropout(dP.drop))