// * Backing store for tensor memory has an associated deallocation
//   function.  This deallocation function will point to client code
//   for tensors populated by the client.  So the client can do things
//   like shadowing a numpy array.
// * We do not provide TF_OK since it is not strictly necessary and we
//   are not optimizing for convenience.
// * We make assumption that one session has one graph.  This should be

/// type or the constness of a parameter, etc.) results in an ABI breakage.
/// Thus, we should not do any of these types of changes, except, potentially,
/// when we are releasing a new major version of TensorFlow. This is an escape
/// hatch, to be used rarely, preferably only to cleanup these structures.
/// Whenever we do these changes, the ABI number must be increased.
///