Split Tests

Author: JesperDramsch

.github/workflows/test_python.yml

@@ -1,4 +1,4 @@
-name: Python package
+name: Test Python package
 
 on: [push]
 

tests/test_conv.py

@@ -0,0 +1,48 @@
+import unittest
+
+from tensorflow.keras.layers import Input, MaxPooling2D, Dense
+from tensorflow.keras.models import Model, Sequential
+import tensorflow as tf
+
+import numpy as np
+import complexnn as conn
+
+
+class TestConvMethods(unittest.TestCase):
+    """Test Conv methods class"""
+
+    def test_conv_outputs_forward(self):
+        """Test computed shape of forward convolution output"""
+        layer = conn.ComplexConv2D(filters=4, kernel_size=3, strides=2, padding="same", transposed=False)
+        input_shape = (None, 128, 128, 2)
+        true = (None, 64, 64, 8)
+        calc = layer.compute_output_shape(input_shape)
+        self.assertEqual(true, calc)
+
+    def test_outputs_transpose(self):
+        """Test computed shape of transposed convolution output"""
+        layer = conn.ComplexConv2D(filters=2, kernel_size=3, strides=2, padding="same", transposed=True)
+        input_shape = (None, 64, 64, 4)
+        true = (None, 128, 128, 4)
+        calc = layer.compute_output_shape(input_shape)
+        self.assertEqual(true, calc)
+
+    def test_conv2D_forward(self):
+        """Test shape of model output, forward"""
+        inputs = Input(shape=(128, 128, 2))
+        outputs = conn.ComplexConv2D(filters=4, kernel_size=3, strides=2, padding="same", transposed=False)(inputs)
+        model = Model(inputs=inputs, outputs=outputs)
+        true = (None, 64, 64, 8)
+        calc = model.output_shape
+        self.assertEqual(true, calc)
+
+    def test_conv2Dtranspose(self):
+        """Test shape of model output, transposed"""
+        inputs = Input(shape=(64, 64, 20))  # = 10 CDN filters
+        outputs = conn.ComplexConv2D(
+            filters=2, kernel_size=3, strides=2, padding="same", transposed=True  # = 4 Keras filters
+        )(inputs)
+        model = Model(inputs=inputs, outputs=outputs)
+        true = (None, 128, 128, 4)
+        calc = model.output_shape
+        self.assertEqual(true, calc)

tests/test_dense.py

@@ -0,0 +1,28 @@
+import unittest
+
+from tensorflow.keras.layers import Input, MaxPooling2D, Dense
+from tensorflow.keras.models import Model, Sequential
+import tensorflow as tf
+
+import numpy as np
+import complexnn as conn
+
+
+class TestDenseMethods(unittest.TestCase):
+    """Test Dense layer"""
+
+    def test_outputs_dense(self):
+        """Test computed shape of dense layer output"""
+        layer = conn.ComplexDense(units=16, activation="relu")
+        input_shape = (None, 8)
+        true = (None, 16 * 2)
+        calc = layer.compute_output_shape(input_shape)
+        self.assertEqual(true, calc)
+
+    def test_outputs_dense(self):
+        """Test computed shape of dense layer output"""
+        layer = conn.ComplexDense(units=16, activation="relu")
+        input_shape = (None, 8)
+        true = (None, 16 * 2)
+        calc = layer.compute_output_shape(input_shape)
+        self.assertEqual(true, calc)

tests/test_readme.py

@@ -0,0 +1,21 @@
+import unittest
+
+from tensorflow.keras.layers import Input, MaxPooling2D, Dense
+from tensorflow.keras.models import Model, Sequential
+import tensorflow as tf
+
+import numpy as np
+import complexnn as conn
+
+
+class TestDNCMethods(unittest.TestCase):
+    """Unit test class"""
+
+    def test_github_example(self):
+        # example from repository https://github.com/JesperDramsch/keras-complex/blob/master/README.md page
+        model = tf.keras.models.Sequential()
+        model.add(conn.conv.ComplexConv2D(32, (3, 3), activation="relu", padding="same", input_shape=(28, 28, 2)))
+        model.add(conn.bn.ComplexBatchNormalization())
+        model.add(MaxPooling2D((2, 2), padding="same"))
+        model.compile(optimizer=tf.keras.optimizers.Adam(), loss="mse")
+        model.summary()

tests/test_train.py

@@ -0,0 +1,49 @@
+import unittest
+
+from tensorflow.keras.layers import Input, MaxPooling2D, Dense
+from tensorflow.keras.models import Model, Sequential
+import tensorflow as tf
+
+import numpy as np
+import complexnn as conn
+
+
+class TestTrainingRuns(unittest.TestCase):
+    """Unit test class"""
+
+    def test_train_transpose(self):
+        """Train using Conv2DTranspose"""
+        x = np.random.randn(64 * 64).reshape((64, 64))
+        y = np.random.randn(64 * 64).reshape((64, 64))
+        X = np.stack((x, y), -1)
+        X = np.expand_dims(X, 0)
+        Y = X
+        inputs = Input(shape=(64, 64, 2))
+        conv1 = conn.ComplexConv2D(
+            filters=2, kernel_size=3, strides=2, padding="same", transposed=False  # = 4 Keras filters
+        )(inputs)
+        outputs = conn.ComplexConv2D(
+            filters=1, kernel_size=3, strides=2, padding="same", transposed=True  # = 2 Keras filters => 1 complex layer
+        )(conv1)
+        model = Model(inputs=inputs, outputs=outputs)
+        model.compile(optimizer="adam", loss="mean_squared_error", metrics=["accuracy"])
+        model.fit(X, Y, batch_size=1, epochs=10)
+
+    def test_train_dense(self):
+        inputs = 28
+        outputs = 128
+        # build a sequential complex dense model
+        model = Sequential(name="complex")
+        model.add(conn.ComplexDense(32, activation="relu", input_shape=(inputs * 2,)))
+        model.add(conn.ComplexBN())
+        model.add(conn.ComplexDense(64, activation="relu"))
+        model.add(conn.ComplexBN())
+        model.add(Dense(128, activation="sigmoid"))
+        model.compile(optimizer="adam", loss="mse")
+        model.summary()
+        # create some random data
+        re = np.random.randn(inputs)
+        im = np.random.randn(inputs)
+        X = np.expand_dims(np.concatenate((re, im), -1), 0)
+        Y = np.expand_dims(np.random.randn(outputs), 0)
+        model.fit(X, Y, batch_size=1, epochs=10)