Fixed small bug with the shape

bob/learn/tensorflow/script/style_transfer.py

@@ -168,10 +168,10 @@ def style_transfer(content_image_path, output_path, style_image_paths,
     """
 
     # Reading and converting to the tensorflow format
-    content_image = bob.io.image.to_matplotlib(bob.io.base.load(content_image_path))
+    content_image = bob.io.image.to_matplotlib(bob.io.base.load(content_image_path)).astype("float32")
     style_images = []
     for path in style_image_paths:
-        image = bob.io.image.to_matplotlib(bob.io.base.load(path))
+        image = bob.io.image.to_matplotlib(bob.io.base.load(path)).astype("float32")
         style_images.append(numpy.reshape(image, wise_shape(image.shape)))
 
     # Reshaping to NxWxHxC

bob/learn/tensorflow/style_transfer/neural_transfer.py

@@ -26,12 +26,11 @@ def compute_features(input_image, architecture, checkpoint_dir, target_end_point
        Dictionary containing the end point tensors
 
     """
-
     input_pl = tf.placeholder('float32', shape=(1, input_image.shape[1],
                                                    input_image.shape[2],
                                                    input_image.shape[3]))
     # TODO: Think on how abstract this normalization operation
-    _, end_points = architecture(tf.stack([tf.image.per_image_standardization(i) for i in tf.unstack(input_image)]), mode=tf.estimator.ModeKeys.PREDICT, trainable_variables=None)
+    _, end_points = architecture(tf.stack([tf.image.per_image_standardization(i) for i in tf.unstack(input_pl)]), mode=tf.estimator.ModeKeys.PREDICT, trainable_variables=None)
 
     with tf.Session() as sess:
         # Restoring the checkpoint for the given architecture