update demo of security defense related to docs

examples/model_security/model_defenses/mnist_defense_nad.py

@@ -64,17 +64,10 @@ def test_nad_method():
 
     # 3. get accuracy of test data on original model
     net.set_train(False)
-    acc_list = []
-    batchs = inputs.shape[0] // batch_size
-    for i in range(batchs):
-        batch_inputs = inputs[i*batch_size : (i + 1)*batch_size]
-        batch_labels = labels[i*batch_size : (i + 1)*batch_size]
-        logits = net(Tensor(batch_inputs)).asnumpy()
-        label_pred = np.argmax(logits, axis=1)
-        acc_list.append(np.mean(batch_labels == label_pred))
-
-    LOGGER.info(TAG, 'accuracy of TEST data on original model is : %s',
-                np.mean(acc_list))
+    logits = net(Tensor(inputs)).asnumpy()
+    label_pred = np.argmax(logits, axis=1)
+    acc = np.mean(labels == label_pred)
+    LOGGER.info(TAG, 'accuracy of TEST data on original model is : %s', acc)
 
     # 4. get adv of test data
     attack = FastGradientSignMethod(net, eps=0.3, loss_fn=loss)
@@ -82,17 +75,11 @@ def test_nad_method():
     LOGGER.info(TAG, 'adv_data.shape is : %s', adv_data.shape)
 
     # 5. get accuracy of adv data on original model
-    acc_list = []
-    batchs = adv_data.shape[0] // batch_size
-    for i in range(batchs):
-        batch_inputs = adv_data[i*batch_size : (i + 1)*batch_size]
-        batch_labels = labels[i*batch_size : (i + 1)*batch_size]
-        logits = net(Tensor(batch_inputs)).asnumpy()
-        label_pred = np.argmax(logits, axis=1)
-        acc_list.append(np.mean(batch_labels == label_pred))
-
-    LOGGER.info(TAG, 'accuracy of adv data on original model is : %s',
-                np.mean(acc_list))
+    net.set_train(False)
+    logits = net(Tensor(adv_data)).asnumpy()
+    label_pred = np.argmax(logits, axis=1)
+    acc = np.mean(labels == label_pred)
+    LOGGER.info(TAG, 'accuracy of adv data on original model is : %s', acc)
 
     # 6. defense
     ds_train = generate_mnist_dataset(os.path.join(mnist_path, "train"),
@@ -111,30 +98,16 @@ def test_nad_method():
 
     # 7. get accuracy of test data on defensed model
     net.set_train(False)
-    acc_list = []
-    batchs = inputs.shape[0] // batch_size
-    for i in range(batchs):
-        batch_inputs = inputs[i*batch_size : (i + 1)*batch_size]
-        batch_labels = labels[i*batch_size : (i + 1)*batch_size]
-        logits = net(Tensor(batch_inputs)).asnumpy()
-        label_pred = np.argmax(logits, axis=1)
-        acc_list.append(np.mean(batch_labels == label_pred))
-
-    LOGGER.info(TAG, 'accuracy of TEST data on defensed model is : %s',
-                np.mean(acc_list))
+    logits = net(Tensor(inputs)).asnumpy()
+    label_pred = np.argmax(logits, axis=1)
+    acc = np.mean(labels == label_pred)
+    LOGGER.info(TAG, 'accuracy of TEST data on defensed model is : %s', acc)
 
     # 8. get accuracy of adv data on defensed model
-    acc_list = []
-    batchs = adv_data.shape[0] // batch_size
-    for i in range(batchs):
-        batch_inputs = adv_data[i*batch_size : (i + 1)*batch_size]
-        batch_labels = labels[i*batch_size : (i + 1)*batch_size]
-        logits = net(Tensor(batch_inputs)).asnumpy()
-        label_pred = np.argmax(logits, axis=1)
-        acc_list.append(np.mean(batch_labels == label_pred))
-
-    LOGGER.info(TAG, 'accuracy of adv data on defensed model is : %s',
-                np.mean(acc_list))
+    logits = net(Tensor(adv_data)).asnumpy()
+    label_pred = np.argmax(logits, axis=1)
+    acc = np.mean(labels == label_pred)
+    LOGGER.info(TAG, 'accuracy of adv data on defensed model is : %s', acc)
 
 
 if __name__ == '__main__':