Merge pull request #223 from jonasrauber/backward_fixes

added backward method support to CompositeModel wrapper

foolbox/models/base.py

@@ -246,7 +246,7 @@ def gradient(self, image, label):
         _, gradient = self.predictions_and_gradient(image, label)
         return gradient
 
-    # TODO: make this an abstract method once support is added to all models
+    @abstractmethod
     def backward(self, gradient, image):
         """Backpropagates the gradient of some loss w.r.t. the logits
         through the network and returns the gradient of that loss w.r.t

foolbox/models/mxnet_gluon.py

@@ -88,3 +88,8 @@ def _loss_fn(self, image, label):
             loss = mx.nd.softmax_cross_entropy(logits, label)
             loss.backward()
         return loss.asnumpy()
+
+    def backward(self, gradient, image):  # pragma: no cover
+        # TODO: backward functionality has not yet been implemented
+        # for MXNetGluonModel
+        raise NotImplementedError

foolbox/models/wrappers.py

@@ -154,6 +154,9 @@ def predictions_and_gradient(self, image, label):
     def gradient(self, image, label):
         return self.backward_model.gradient(image, label)
 
+    def backward(self, gradient, image):
+        return self.backward_model.backward(gradient, image)
+
     def __enter__(self):
         assert self.forward_model.__enter__() == self.forward_model
         assert self.backward_model.__enter__() == self.backward_model

foolbox/tests/test_model_wrappers.py

@@ -39,6 +39,8 @@ def test_diff_wrapper(bn_model, bn_image, bn_label):
 
 
 def test_composite_model(gl_bn_model, bn_model, bn_image, bn_label):
+    num_classes = 10
+    test_grad = np.random.rand(num_classes).astype(np.float32)
     model = CompositeModel(gl_bn_model, bn_model)
     with model:
         assert gl_bn_model.num_classes() == model.num_classes()
@@ -48,6 +50,9 @@ def test_composite_model(gl_bn_model, bn_model, bn_image, bn_label):
         assert np.all(
             bn_model.gradient(bn_image, bn_label) ==
             model.gradient(bn_image, bn_label))
+        assert np.all(
+            bn_model.backward(test_grad, bn_image) ==
+            model.backward(test_grad, bn_image))
         assert np.all(
             gl_bn_model.predictions(bn_image) ==
             model.predictions_and_gradient(bn_image, bn_label)[0])

foolbox/tests/test_models.py

@@ -44,6 +44,9 @@ def num_classes(self):
         def predictions_and_gradient(self, image, label):
             return 'predictions', 'gradient'
 
+        def backward(self, gradient, image):
+            return image
+
     model = TestModel(bounds=(0, 1), channel_axis=1)
 
     image = np.ones((28, 28, 1), dtype=np.float32)