Use named constant num_classes instead of literals (keras-team#8129)

tests/integration_tests/test_vector_data_tasks.py

@@ -7,6 +7,8 @@
 import keras
 from keras.utils.np_utils import to_categorical
 
+num_classes = 2
+
 
 @keras_test
 def test_vector_classification():
@@ -18,7 +20,7 @@ def test_vector_classification():
                                                          num_test=200,
                                                          input_shape=(20,),
                                                          classification=True,
-                                                         num_classes=2)
+                                                         num_classes=num_classes)
     y_train = to_categorical(y_train)
     y_test = to_categorical(y_test)
 
@@ -27,7 +29,7 @@ def test_vector_classification():
         layers.Dense(16, input_shape=(x_train.shape[-1],), activation='relu'),
         layers.Dense(8),
         layers.Activation('relu'),
-        layers.Dense(y_train.shape[-1], activation='softmax')
+        layers.Dense(num_classes, activation='softmax')
     ])
     model.compile(loss='categorical_crossentropy',
                   optimizer='rmsprop',
@@ -47,13 +49,13 @@ def test_vector_classification_functional():
                                                          num_test=200,
                                                          input_shape=(20,),
                                                          classification=True,
-                                                         num_classes=2)
+                                                         num_classes=num_classes)
     # Test with functional API
     inputs = layers.Input(shape=(x_train.shape[-1],))
     x = layers.Dense(16, activation=keras.activations.relu)(inputs)
     x = layers.Dense(8)(x)
     x = layers.Activation('relu')(x)
-    outputs = layers.Dense(y_train.shape[-1], activation='softmax')(x)
+    outputs = layers.Dense(num_classes, activation='softmax')(x)
     model = keras.models.Model(inputs, outputs)
     model.compile(loss=keras.losses.sparse_categorical_crossentropy,
                   optimizer=keras.optimizers.RMSprop(),
@@ -73,12 +75,12 @@ def test_vector_regression():
     (x_train, y_train), (x_test, y_test) = get_test_data(num_train=500,
                                                          num_test=200,
                                                          input_shape=(20,),
-                                                         output_shape=(2,),
+                                                         output_shape=(num_classes,),
                                                          classification=False)
 
     model = Sequential([
         layers.Dense(16, input_shape=(x_train.shape[-1],), activation='tanh'),
-        layers.Dense(y_train.shape[-1])
+        layers.Dense(num_classes)
     ])
 
     model.compile(loss='hinge', optimizer='adagrad')

tests/keras/legacy/models_test.py

@@ -40,7 +40,7 @@ def _get_test_data():
                                                          num_test=test_samples,
                                                          input_shape=(input_dim,),
                                                          classification=True,
-                                                         num_classes=4)
+                                                         num_classes=num_classes)
     y_test = np_utils.to_categorical(y_test)
     y_train = np_utils.to_categorical(y_train)
     return (x_train, y_train), (x_test, y_test)

tests/keras/optimizers_test.py

@@ -11,14 +11,16 @@
 from keras.utils.np_utils import to_categorical
 from keras import backend as K
 
+num_classes = 2
+
 
 def get_test_data():
     np.random.seed(1337)
     (x_train, y_train), _ = test_utils.get_test_data(num_train=1000,
                                                      num_test=200,
                                                      input_shape=(10,),
                                                      classification=True,
-                                                     num_classes=2)
+                                                     num_classes=num_classes)
     y_train = to_categorical(y_train)
     return x_train, y_train
 
@@ -123,9 +125,9 @@ def test_tfoptimizer():
     from tensorflow import train
     optimizer = optimizers.TFOptimizer(train.AdamOptimizer())
     model = Sequential()
-    model.add(Dense(2, input_shape=(3,), kernel_constraint=constraints.MaxNorm(1)))
+    model.add(Dense(num_classes, input_shape=(3,), kernel_constraint=constraints.MaxNorm(1)))
     model.compile(loss='mean_squared_error', optimizer=optimizer)
-    model.fit(np.random.random((5, 3)), np.random.random((5, 2)),
+    model.fit(np.random.random((5, 3)), np.random.random((5, num_classes)),
               epochs=1, batch_size=5, verbose=0)
     # not supported
     with pytest.raises(NotImplementedError):

tests/keras/test_callbacks.py

@@ -634,7 +634,7 @@ def test_TensorBoard_convnet(tmpdir):
                                                          num_test=200,
                                                          input_shape=input_shape,
                                                          classification=True,
-                                                         num_classes=4)
+                                                         num_classes=num_classes)
     y_train = np_utils.to_categorical(y_train)
     y_test = np_utils.to_categorical(y_test)
 
@@ -646,7 +646,7 @@ def test_TensorBoard_convnet(tmpdir):
         Conv2D(filters=4, kernel_size=(3, 3),
                activation='relu', padding='same'),
         GlobalAveragePooling2D(),
-        Dense(y_test.shape[-1], activation='softmax')
+        Dense(num_classes, activation='softmax')
     ])
     model.compile(loss='categorical_crossentropy',
                   optimizer='rmsprop',

tests/keras/test_sequential_model.py

@@ -60,7 +60,7 @@ def _get_test_data():
                                                          num_test=test_samples,
                                                          input_shape=(input_dim,),
                                                          classification=True,
-                                                         num_classes=4)
+                                                         num_classes=num_classes)
     y_test = np_utils.to_categorical(y_test)
     y_train = np_utils.to_categorical(y_train)
     return (x_train, y_train), (x_test, y_test)
@@ -276,8 +276,8 @@ def test_clone_functional_model():
 
     input_a = keras.Input(shape=(4,))
     input_b = keras.Input(shape=(4,))
-    dense_1 = keras.layers.Dense(4,)
-    dense_2 = keras.layers.Dense(4,)
+    dense_1 = keras.layers.Dense(4)
+    dense_2 = keras.layers.Dense(4)
 
     x_a = dense_1(input_a)
     x_a = keras.layers.Dropout(0.5)(x_a)