refactor: add loading of the model, saved as protobuf, and make predi…

…ction on single image, using it

svnh_semi_supervised_model_loaded_test.py

@@ -27,7 +27,10 @@ def load_test_images():
     return test_images, test_labels
 
 
-def main(_):
+def load_and_predict_with_checkpoints():
+    '''
+    Loads saved model checkpoints and make prediction on test images
+    '''
     # load test images and labels
     test_images, test_labels = load_test_images()
 
@@ -59,5 +62,37 @@ def main(_):
         print("Predicted classes: {}".format(pred_class))
 
 
+def load_and_predict_with_saved_model():
+    '''
+    Loads saved as protobuf model and make prediction on a single image
+    '''
+    with tf.Session(graph=tf.Graph()) as sess:
+        # restore save model
+        export_dir = './gan-export/1'
+        model = tf.saved_model.loader.load(sess, [tf.saved_model.tag_constants.SERVING], export_dir)
+        # print(model)
+        loaded_graph = tf.get_default_graph()
+
+        # get necessary tensors by name
+        input_tensor_name = model.signature_def['predict_images'].inputs['images'].name
+        input_tensor = loaded_graph.get_tensor_by_name(input_tensor_name)
+        output_tensor_name = model.signature_def['predict_images'].outputs['scores'].name
+        output_tensor = loaded_graph.get_tensor_by_name(output_tensor_name)
+
+        # make prediction
+        image_file_name = './svnh_test_images/image_3.jpg'
+        with open(image_file_name, 'rb') as f:
+            image = f.read()
+            scores = sess.run(output_tensor, {input_tensor: [image]})
+
+        # print results
+        print("Scores: {}".format(scores))
+
+
+def main(_):
+    # load_and_predict_with_checkpoints()
+    load_and_predict_with_saved_model()
+
+
 if __name__ == '__main__':
     tf.app.run()