Week 2 - Pipeline CNN, eksperimen, laporan akhir

src/models.py

@@ -80,3 +80,38 @@ def create_cnn_mnist_model(input_shape: tuple = (28, 28, 1), num_classes: int =
     )
     return model
 
+
+def create_deep_cnn_mnist_model(input_shape: tuple = (28, 28, 1), num_classes: int = 10) -> keras.Model:
+    """
+    Membangun model CNN yang lebih dalam dengan 3 blok Conv2D + MaxPooling2D.
+    Digunakan untuk Tugas 2 (Eksplorasi Kedalaman Arsitektur CNN).
+    """
+    model = keras.Sequential(
+        name="cnn_mnist_3blocks",
+        layers=[
+            layers.Input(shape=input_shape),
+            # Block 1
+            layers.Conv2D(32, (3, 3), activation="relu", padding="same"),
+            layers.MaxPooling2D((2, 2)),
+            # Block 2
+            layers.Conv2D(64, (3, 3), activation="relu", padding="same"),
+            layers.MaxPooling2D((2, 2)),
+            # Block 3 (Blok tambahan)
+            layers.Conv2D(128, (3, 3), activation="relu", padding="same"),
+            layers.MaxPooling2D((2, 2)),
+            # Classifier
+            layers.Flatten(),
+            layers.Dense(64, activation="relu"),
+            layers.Dropout(0.3),
+            layers.Dense(num_classes, activation="softmax"),
+        ]
+    )
+
+    model.compile(
+        optimizer="adam",
+        loss="sparse_categorical_crossentropy",
+        metrics=["accuracy"]
+    )
+    return model
+
+