Fix bug concerning computation of gradients inside a CondContext.

Change: 117260468

tensorflow/python/ops/control_flow_grad.py

@@ -142,7 +142,7 @@ def _ExitGrad(_, grad):
   # pylint: enable=protected-access
   if not grad_ctxt.back_prop:
     # The flag `back_prop` is set by users to suppress gradient
-    # computation for this loop. If the flag `back_prop` is true,
+    # computation for this loop. If the attribute `back_prop` is false,
     # no gradient computation.
     return None
   grad_ctxt.AddName(grad.name)
@@ -184,7 +184,7 @@ def _EnterGrad(op, grad):
   grad_ctxt = graph._get_control_flow_context()
   # pylint: enable=protected-access
   if not grad_ctxt.back_prop:
-    # If the flag `back_prop` is true, no gradient computation.
+    # If the attribute `back_prop` is true, no gradient computation.
     return grad
   if op.get_attr("is_constant"):
     # Add a gradient accumulator for each loop invariant.

tensorflow/python/ops/control_flow_ops.py

@@ -909,6 +909,14 @@ class ControlFlowContext(object):
     """Return the context containing this context."""
     return self._outer_context
 
+  @property
+  def grad_state(self):
+    raise NotImplementedError("Abstract method")
+
+  @property
+  def back_prop(self):
+    raise NotImplementedError("Abstract method")
+
   def AddName(self, name):
     self._values.add(name)
 
@@ -979,6 +987,18 @@ class CondContext(ControlFlowContext):
   def branch(self):
     return self._branch
 
+  @property
+  def grad_state(self):
+    if self.GetWhileContext():
+      return self.GetWhileContext().grad_state
+    return None
+
+  @property
+  def back_prop(self):
+    if self.GetWhileContext():
+      self.GetWhileContext().back_prop
+    return False
+
   def AddValue(self, val):
     """Add `val` to the current context and its outer context recursively."""
     result = val