Use NDShape.unknown instead of None

bindings/python/cntk/cntk_py.i

@@ -387,13 +387,11 @@ fail:
 %typecheck(1000) CNTK::NDShape const &, CNTK::NDShape {
     // '1000' is the typecheck precedence code. It means: check after basic
     // types, but before arrays. See: http://www.swig.org/Doc1.3/Typemaps.html#Typemaps_overloading
-    $1 = (($input == Py_None) || PyTuple_Check($input)) ? 1 : 0;
+    $1 = PyTuple_Check($input) ? 1 : 0;
 }
 
 %typemap(in) CNTK::NDShape const & {
-     if ($input == Py_None) {
-        $1 = new CNTK::NDShape(CNTK::NDShape::Unknown);
-     } else if (PyTuple_Check($input)) {
+     if (PyTuple_Check($input)) {
         size_t rank = PyTuple_Size($input);
         std::vector<size_t> dimensions(rank);
         for (size_t i=0; i<rank; i++)

bindings/python/cntk/layers.py

@@ -177,7 +177,7 @@ def Convolution(filter_shape,        # e.g. (3,3)
 #   AveragePooling and GlobalAveragePooling
 #
 # Setting the filter_shape to None, mean global pooling.
-from cntk.cntk_py import PoolingType_Max, PoolingType_Average
+from cntk.cntk_py import PoolingType_Max, PoolingType_Average, NDShape
 def Pooling(op,      # PoolingType_Max or _Average
             filter_shape,  # e.g. (3,3)
             strides=1,
@@ -207,11 +207,11 @@ def AveragePooling(filter_shape,  # e.g. (3,3)
 
 # GlobalMaxPooling
 def GlobalMaxPooling():
-    return Pooling(PoolingType_Max, None, pad=False)
+    return Pooling(PoolingType_Max, NDShape.unknown.dimensions(), pad=False)
 
 # GlobalAveragePooling
 def GlobalAveragePooling():
-    return Pooling(PoolingType_Average, None, pad=False)
+    return Pooling(PoolingType_Average, NDShape.unknown.dimensions(), pad=False)
 
 # Recurrence() -- run a block recurrently over a time sequence
 def Recurrence(over, go_backwards=False, initial_state=initial_state_default_or_None):

bindings/python/cntk/utils/__init__.py

@@ -183,12 +183,8 @@ def get_temp_filename(directory=None):
 
 def sanitize_shape(shape):
     """
-    If shape is scalar, it creates a tuple out of it. If the shape is None, then return it as it is, it will be mapped as NDShape::Unknown.
+    If shape is scalar, it creates a tuple out of it.
     """
-
-    # Unknown shape
-    if shape is None:
-        return shape
     return _as_tuple(shape)