Updated documentation to use .meta files

example/tensorflow/code_template/tensorflow_sample.py

@@ -11,9 +11,9 @@
 # Read the protobuf text and build a tf.GraphDef
 with open(model_file_name, 'r') as model_file:
     model_protobuf = text_format.Parse(model_file.read(),
-                                       tf.GraphDef())
+                                       tf.MetaGraphDef())
 
 # Import the GraphDef built above into the default graph
-tf.import_graph_def(model_protobuf)
+tf.train.import_meta_graph(model_protobuf)
 
 # You can now add operations on top of the imported graph

tutorials/tensorflow_prototxt_usage.md

@@ -6,34 +6,34 @@ In order to export a Tensorflow model from Fabrik:
 <img src="https://raw.githubusercontent.com/Cloud-CV/Fabrik/master/tutorials/exportbutton.png">
 
 2. A drop-down list should appear. Select Tensorflow.
-    * This should download a pbtxt file to your computer.  
+    * This should download a ```.meta``` file to your computer.  
     <img src="https://raw.githubusercontent.com/Cloud-CV/Fabrik/master/tutorials/export_tensorflow.png">
 
-3. Rename the file to ```model.pbtxt```.
+3. Rename the file to ```model.meta```.
 
-4. Load the model from the ProtoBuf file using the following code:
+4. Load the model from ```model.meta``` using the following code:
 
     ```
     import tensorflow as tf
     from google.protobuf import text_format
 
     # read the graphdef from the model file
-    with open('model.pbtxt', 'r') as model_file:
+    with open('model.meta', 'r') as model_file:
 	    model_protobuf = text_format(model_file.read(),
-		    	                 tf.Graphdef())
+		    	                 tf.MetaGraphDef())
     
     # import the graphdef into the default graph
-    tf.import_graph_def(model_protobuf)
+    tf.train.import_meta_graph(model_protobuf)
     ```
 
 ### Code template
 
-[The code template](../example/tensorflow/code_template/tensorflow_sample.py) loads the model from a pbtxt file into the default graph. Additional operations like layers and optimizers can be then built onto the graph as required.
+[The code template](../example/tensorflow/code_template/tensorflow_sample.py) loads the model from a ```.meta``` file into the default graph. Additional operations like layers and optimizers can be then built onto the graph as required.
 
 To run the code, run:
 
 ```
-python tensorflow_sample.py model.pbtxt
+python tensorflow_sample.py model.meta
 ```
 
-Replace ```model.pbtxt``` with the model file that you want to use.
+Replace ```model.meta``` with the model file that you want to use.