testing::CreateTensorHandle(ctx, DT_FLOAT, {2, 4}, value)) {}
};

FunctionDef FuncDefWithNumInputsOutputs(int num_inputs, int num_outputs) {
  FunctionDef func;
  OpDef* signature = func.mutable_signature();
  for (int i = 0; i < num_inputs; ++i) {
    signature->add_input_arg();
  }
  for (int i = 0; i < num_outputs; ++i) {
    signature->add_output_arg();
  }
  return func;
}

                                        &rewritten_fdef));
  const auto new_function_name =
      fld->UniqueFunctionName(absl::StrCat(n->type_string(), "_lifted_arg_"));
  rewritten_fdef.mutable_signature()->set_name(new_function_name);
  TF_RETURN_IF_ERROR(fld->AddFunctionDef(rewritten_fdef));

  // We need to recreate the node. Otherwise TF will not know n->num_inputs()
  // has increased.

  }

  auto stateful_string = mlir::TF::TensorFlowDialect::GetStatefulAttrName();
  if (auto attr = function->getAttrOfType<mlir::UnitAttr>(stateful_string)) {
    func_def.mutable_signature()->set_is_stateful(true);
  }

  // Ignore the gradient and is_stateful attribute on the function as they have
  // been handled above. Ignore the entry func attribute as it is an MLIR

  }

  auto stateful_string = mlir::TF::TensorFlowDialect::GetStatefulAttrName();
  if (auto attr = function->getAttrOfType<mlir::UnitAttr>(stateful_string)) {
    func_def.mutable_signature()->set_is_stateful(true);
  }

  // Ignore the gradient and is_stateful attribute on the function as they have
  // been handled above. Ignore the entry func attribute as it is an MLIR

      FunctionDefHelper::Define("UncompilableFn", {"n_a:float"}, {"n_c:float"},
                                {}, {{{"n_c"}, "UncompilableUnary", {"n_a"}}});
  FunctionDef noinline = compilable;
  noinline.mutable_signature()->set_name("NoInlineFn");
  AddAttr("_noinline", static_cast<bool>(true), noinline.mutable_attr());

  FunctionDefLibrary flib;
  *flib.add_function() = compilable;
  *flib.add_function() = uncompilable;