diff --git a/au-zone/readme.md b/au-zone/readme.md
index 39ed1e2..21ff227 100644
--- a/au-zone/readme.md
+++ b/au-zone/readme.md
@@ -70,58 +70,3 @@ Mandatory Arguments:
 	--input-model
 	--data-model (Caffe2)
 	--value-info (Caffe2)
-    
-
-## Project Architecture
-
-For a given conversion task, convert.py loads an importer and exporter and runs the conversion by importing into a NNEF 
-graph representation (based on NetworkX) and exports to the proper format. 
-
-### Framework support
-Each supported framework implements the import and export functions. Framework implementations are self contained into 
-their directory ('tensorflow', caffe2', 'common' for NNEF, etc). Existing framework implementations should be used as 
-references for supporting additional frameworks (once import-export functions are implemented, the implementer just needs 
-to support it in convert.py).
-
-### NNEF representation
-Importers/exporters are based on NNEF Fragment and NNEF Graph objects (from "common" directory). 
-
-The importers are parsing data from the input format and are reinterpreting it as NNEF definitions to create a NNEF 
-Fragment for each operation (layer).The importers then use the Fragments created to create a NNEF Graph representation.
- 
-The exporters are using a NNEF Graph object to export either to NNEF format or another framework. 
-
-Here is a simplified example showing the NNEF Fragments and NNEF Graph APIs: 
-
-    # Creating NNEF 'External' fragment 
-    input = op.External(shape=in_shape, _uid='input')
-
-    # Creating NNEF 'Variable' fragment
-    var1 = op.Variable(label='vars/var1', shape=in_shape, _np_tensor = np_array, _uid='var1')
-
-    # Creating NNEF 'Add' fragment
-    output = op.Add(x=input, y=var1, _uid='output')
-
-    # Creating NNEF Graph 
-    model = NNEFGraph('test_model', 
-                      input, 
-                      output)
-                      
-    # Running inference (assuming np_array has previously been set...)
-    input_data = {'input':input_tensor}
-    output_tensors = ['output']
-    model_output = nnef_model.run(output_tensors, input_data)
-
-    # 'ground_truth' holds the inference result as a numpy array
-    ground_truth = model_output['output']
-
-
-## Unit Tests
-
-The project contains 2 levels of unit tests. 
-
-- Syntax testing: Verifies that a conversion from a framework to NNEF and back to the original framework 
-still provides a valid model (in the original framework). Those tests are implemented for each of the NNEF primitives, for each supported framework.
-
-- Model testing: Verifies that the graph result of networks are equivalent before and after converting them to NNEF 
-and back to their original format. Those tests are implemented for a few reference networks (those provided by the TensorFlow and Caffe2 Model Zoos).