Try to make baseline model less prone to overfitting

Training history showed that (proper) val metrics (evaluated on a
different synthesis tree) behave better with this model version. A
clear val_loss minimum is now seen (and I'm overtraining the model).

ml/synthesis/src/components/model_generation/models.py

@@ -5,16 +5,16 @@ def create_baseline_model(input_shape) -> Model:
     model = Sequential(
         [
             layers.InputLayer(input_shape=input_shape),
-            layers.Dense(128, activation="relu"),
-            layers.Dense(64, activation="relu"),
-            layers.Dense(64, activation="relu"),
-            layers.Dense(32, activation="relu"),
+            layers.Dense(64, activation="relu", kernel_regularizer="l2"),
+            layers.Dropout(0.5),
+            layers.Dense(16, activation="relu"),
+            layers.Dense(16, activation="relu"),
             layers.Dense(1),
         ],
     )
 
     model.compile(
-        optimizer=optimizers.Adam(learning_rate=1e-3),
+        optimizer=optimizers.Adam(learning_rate=1e-4),
         loss="mse",
         metrics=["mae"],
     )