New test cases

bob/learn/tensorflow/test/test_cnn_pretrained_model.py

@@ -256,4 +256,5 @@ def test_siamese_cnn_pretrained():
 
     del trainer
     tf.reset_default_graph()
-    assert len(tf.global_variables())==0    
+    assert len(tf.global_variables())==0
+

bob/learn/tensorflow/test/test_cnn_scratch.py

@@ -28,7 +28,10 @@ slim = tf.contrib.slim
 
 def scratch_network(train_data_shuffler, reuse=False):
 
-    inputs = train_data_shuffler("data", from_queue=False)
+    if isinstance(train_data_shuffler, tf.Tensor):
+        inputs = train_data_shuffler
+    else:
+        inputs = train_data_shuffler("data", from_queue=False)
 
     # Creating a random network
     initializer = tf.contrib.layers.xavier_initializer(seed=seed)
@@ -150,7 +153,7 @@ def test_cnn_trainer_scratch_tfrecord():
                                          
     graph = scratch_network(train_data_shuffler)
     validation_graph = scratch_network(validation_data_shuffler, reuse=True)
-
+    
     # Setting the placeholders
     # Loss for the softmax
     loss = MeanSoftMaxLoss()
@@ -174,9 +177,27 @@ def test_cnn_trainer_scratch_tfrecord():
     
     trainer.train()
     os.remove(tfrecords_filename)
-    os.remove(tfrecords_filename_val)
+    os.remove(tfrecords_filename_val)    
     assert True
     tf.reset_default_graph()
     del trainer
     assert len(tf.global_variables())==0
+    
+    # Inference. TODO: Wrap this in a package
+    file_name = os.path.join(directory, "model.ckp.meta")
+    images = tf.placeholder(tf.float32, shape=(None, 28, 28, 1))
+    graph = scratch_network(images, reuse=False)
+
+    session = tf.Session()
+    session.run(tf.global_variables_initializer())
+    saver = tf.train.import_meta_graph(file_name, clear_devices=True)
+    saver.restore(session, tf.train.latest_checkpoint(os.path.dirname("./temp/cnn_scratch/")))
+    data = numpy.random.rand(2, 28, 28, 1).astype("float32")
+
+    assert session.run(graph, feed_dict={images: data}).shape == (2, 10)
+
+    tf.reset_default_graph()
+    shutil.rmtree(directory)
+    assert len(tf.global_variables())==0
+