Standardized Testing for All Optimizers

tensorflow_addons/optimizers/tests/standard_test.py

@@ -0,0 +1,49 @@
+import pytest
+import numpy as np
+import tensorflow as tf
+
+from tensorflow_addons import optimizers
+import inspect
+import sys
+
+classes_to_test = ["RectifiedAdam", "LazyAdam"]
+
+
+def discover_classes(module, parent):
+
+    classes = [
+        m[1]
+        for m in inspect.getmembers(module, inspect.isclass)
+        if issubclass(m[1], parent) and m[0] in classes_to_test
+    ]
+
+    return classes
+
+
+@pytest.mark.parametrize(
+    "optimizer", discover_classes(optimizers, tf.keras.optimizers.Optimizer)
+)
+def test_optimizer_minimize(optimizer):
+
+    model = tf.keras.Sequential([tf.keras.Input(shape=[1]), tf.keras.layers.Dense(1)])
+
+    x = np.array(np.ones([1]))
+    y = np.array(np.zeros([1]))
+
+    opt = optimizer()
+    loss = tf.keras.losses.MSE
+
+    model.compile(optimizer=opt, loss=loss)
+
+    history = model.fit(x, y, batch_size=1, epochs=10)
+
+    loss_values = history.history["loss"]
+
+    np.testing.assert_array_less(loss_values[-1], loss_values[0])
+
+
+if __name__ == "__main__":
+    sys.exit(pytest.main(["standard_test.py"]))
+
+    # python -m flake8 tensorflow_addons/optimizers/tests/standard_test.py
+    # python -m black tensorflow_addons/optimizers/tests/standard_test.py