def test_raises_type_error_with_federated_type_at_clients(
self, executor, num_clients):
value, type_signature = executor_test_utils.create_dummy_value_at_clients(
num_clients)
value = self.run_sync(executor.create_value(value, type_signature))
with self.assertRaises(TypeError):
self.run_sync(
executor_utils.compute_intrinsic_federated_broadcast(executor, value))
def test_raises_not_implemented_error_with_intrinsic_def_federated_secure_sum(
self):
executor = create_test_executor()
comp, comp_type = executor_test_utils.create_dummy_intrinsic_def_federated_secure_sum(
)
arg_1, arg_1_type = executor_test_utils.create_dummy_value_at_clients()
arg_2, arg_2_type = executor_test_utils.create_dummy_value_unplaced()
comp = self.run_sync(executor.create_value(comp, comp_type))
arg_1 = self.run_sync(executor.create_value(arg_1, arg_1_type))
arg_2 = self.run_sync(executor.create_value(arg_2, arg_2_type))
args = self.run_sync(executor.create_tuple([arg_1, arg_2]))
with self.assertRaises(NotImplementedError):
self.run_sync(executor.create_call(comp, args))
def test_raises_type_error_with_singleton_tuple(
self,
executor,
num_clients,
):
value, type_signature = executor_test_utils.create_dummy_value_at_clients(
num_clients)
value = self.run_sync(executor.create_value(value, type_signature))
arg = self.run_sync(executor.create_struct([value]))
with self.assertRaises(TypeError):
self.run_sync(
executor_utils.compute_intrinsic_federated_weighted_mean(
executor, arg))
def test_computes_weighted_mean(
self,
executor,
num_clients,
):
value, type_signature = executor_test_utils.create_dummy_value_at_clients(
num_clients)
# Weighted mean computed in Python
expected_result = sum([x**2 for x in value]) / sum(value)
value = self.run_sync(executor.create_value(value, type_signature))
arg = self.run_sync(executor.create_struct([value, value]))
result = self.run_sync(
executor_utils.compute_intrinsic_federated_weighted_mean(executor, arg))
self.assertIsInstance(result, executor_value_base.ExecutorValue)
expected_type = computation_types.at_server(type_signature.member)
self.assertEqual(result.type_signature.compact_representation(),
expected_type.compact_representation())
actual_result = self.run_sync(result.compute())
self.assertEqual(actual_result, expected_result)
class FederatingExecutorCreateTupleTest(executor_test_utils.AsyncTestCase,
parameterized.TestCase):
# pyformat: disable
@parameterized.named_parameters([
('placement_literal',
*executor_test_utils.create_dummy_placement_literal()),
('computation_call',
*executor_test_utils.create_dummy_computation_call()),
('computation_placement',
*executor_test_utils.create_dummy_computation_placement()),
('computation_selection',
*executor_test_utils.create_dummy_computation_selection()),
('computation_tuple',
*executor_test_utils.create_dummy_computation_tuple()),
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
('unplaced_type', *executor_test_utils.create_dummy_value_unplaced()),
])
# pyformat: enable
def test_returns_value_with_elements_value(self, value, type_signature):
executor = create_test_executor()
element = self.run_sync(executor.create_value(value, type_signature))
elements = [element] * 3
type_signature = computation_types.NamedTupleType([type_signature] * 3)
result = self.run_sync(executor.create_tuple(elements))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
# TODO(b/153578410): Some `FederatingExecutorValue` that can can be
# constructed, can not be computed.
# pyformat: disable
@parameterized.named_parameters([
('intrinsic_def_federated_eval_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_eval_at_server(),
*executor_test_utils.create_dummy_computation_tensorflow_constant()),
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic(),
*executor_test_utils.create_dummy_computation_tensorflow_constant()),
])
# pyformat: enable
def test_returns_value_with_elements_fn_and_arg(self, fn, fn_type, arg,
arg_type):
executor = create_test_executor()
fn = self.run_sync(executor.create_value(fn, fn_type))
arg = self.run_sync(executor.create_value(arg, arg_type))
element = self.run_sync(executor.create_call(fn, arg))
elements = [element] * 3
type_signature = computation_types.NamedTupleType([fn_type.result] * 3)
result = self.run_sync(executor.create_tuple(elements))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
actual_result = self.run_sync(result.compute())
expected_result = [self.run_sync(element.compute())] * 3
self.assertCountEqual(actual_result, expected_result)
# pyformat: disable
@parameterized.named_parameters([
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
])
# pyformat: enable
def test_returns_value_with_elements_fn_only(self, fn, fn_type):
executor = create_test_executor()
fn = self.run_sync(executor.create_value(fn, fn_type))
element = self.run_sync(executor.create_call(fn))
elements = [element] * 3
type_signature = computation_types.NamedTupleType([fn_type.result] * 3)
result = self.run_sync(executor.create_tuple(elements))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
actual_result = self.run_sync(result.compute())
expected_result = [self.run_sync(element.compute())] * 3
self.assertCountEqual(actual_result, expected_result)
def test_raises_type_error_with_unembedded_elements(self):
executor = create_test_executor()
element, _ = executor_test_utils.create_dummy_value_unplaced()
elements = [element] * 3
with self.assertRaises(TypeError):
self.run_sync(executor.create_tuple(elements))
class FederatingExecutorCreateCallTest(executor_test_utils.AsyncTestCase,
parameterized.TestCase):
# pyformat: disable
@parameterized.named_parameters([
('intrinsic_def_federated_aggregate',
*executor_test_utils.create_dummy_intrinsic_def_federated_aggregate(),
[
executor_test_utils.create_dummy_value_at_clients(),
executor_test_utils.create_dummy_value_unplaced(),
executor_test_utils.create_dummy_computation_tensorflow_add(),
executor_test_utils.create_dummy_computation_tensorflow_add(),
executor_test_utils.create_dummy_computation_tensorflow_identity(
)
], 43.0),
('intrinsic_def_federated_apply',
*executor_test_utils.create_dummy_intrinsic_def_federated_apply(), [
executor_test_utils.create_dummy_computation_tensorflow_identity(
),
executor_test_utils.create_dummy_value_at_server()
], 10.0),
('intrinsic_def_federated_broadcast',
*executor_test_utils.create_dummy_intrinsic_def_federated_broadcast(),
[executor_test_utils.create_dummy_value_at_server()], 10.0),
('intrinsic_def_federated_collect',
*executor_test_utils.create_dummy_intrinsic_def_federated_collect(),
[executor_test_utils.create_dummy_value_at_clients()],
tf.data.Dataset.from_tensor_slices([10.0, 11.0, 12.0])),
('intrinsic_def_federated_eval_at_clients', *executor_test_utils.
create_dummy_intrinsic_def_federated_eval_at_clients(),
[executor_test_utils.create_dummy_computation_tensorflow_constant()
], [10.0] * 3),
('intrinsic_def_federated_eval_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_eval_at_server(),
[executor_test_utils.create_dummy_computation_tensorflow_constant()
], 10.0),
('intrinsic_def_federated_map',
*executor_test_utils.create_dummy_intrinsic_def_federated_map(), [
executor_test_utils.create_dummy_computation_tensorflow_identity(
),
executor_test_utils.create_dummy_value_at_clients()
], [10.0, 11.0, 12.0]),
('intrinsic_def_federated_map_all_equal', *executor_test_utils.
create_dummy_intrinsic_def_federated_map_all_equal(), [
executor_test_utils.create_dummy_computation_tensorflow_identity(
),
executor_test_utils.create_dummy_value_at_clients_all_equal()
], 10.0),
('intrinsic_def_federated_mean',
*executor_test_utils.create_dummy_intrinsic_def_federated_mean(),
[executor_test_utils.create_dummy_value_at_clients()], 11.0),
('intrinsic_def_federated_sum',
*executor_test_utils.create_dummy_intrinsic_def_federated_sum(),
[executor_test_utils.create_dummy_value_at_clients()], 33.0),
('intrinsic_def_federated_reduce',
*executor_test_utils.create_dummy_intrinsic_def_federated_reduce(), [
executor_test_utils.create_dummy_value_at_clients(),
executor_test_utils.create_dummy_value_unplaced(),
executor_test_utils.create_dummy_computation_tensorflow_add()
], 43.0),
('intrinsic_def_federated_value_at_clients', *executor_test_utils.
create_dummy_intrinsic_def_federated_value_at_clients(),
[executor_test_utils.create_dummy_value_unplaced()], 10.0),
('intrinsic_def_federated_value_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_value_at_server(),
[executor_test_utils.create_dummy_value_unplaced()], 10.0),
('intrinsic_def_federated_weighted_mean', *executor_test_utils.
create_dummy_intrinsic_def_federated_weighted_mean(), [
executor_test_utils.create_dummy_value_at_clients(),
executor_test_utils.create_dummy_value_at_clients()
], 11.060606),
('intrinsic_def_federated_zip_at_clients', *executor_test_utils.
create_dummy_intrinsic_def_federated_zip_at_clients(), [
executor_test_utils.create_dummy_value_at_clients(),
executor_test_utils.create_dummy_value_at_clients()
], [
anonymous_tuple.AnonymousTuple([(None, 10.0), (None, 10.0)]),
anonymous_tuple.AnonymousTuple([(None, 11.0), (None, 11.0)]),
anonymous_tuple.AnonymousTuple([(None, 12.0), (None, 12.0)])
]),
('intrinsic_def_federated_zip_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_zip_at_server(), [
executor_test_utils.create_dummy_value_at_server(),
executor_test_utils.create_dummy_value_at_server()
], anonymous_tuple.AnonymousTuple([(None, 10.0), (None, 10.0)])),
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic(),
[executor_test_utils.create_dummy_computation_tensorflow_constant()
], 10.0),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_identity(),
[executor_test_utils.create_dummy_value_unplaced()], 10.0),
])
# pyformat: enable
def test_returns_value_with_comp_and_arg(self, comp, comp_type, args,
expected_result):
executor = create_test_executor()
comp = self.run_sync(executor.create_value(comp, comp_type))
elements = [self.run_sync(executor.create_value(*x)) for x in args]
if len(elements) > 1:
arg = self.run_sync(executor.create_tuple(elements))
else:
arg = elements[0]
result = self.run_sync(executor.create_call(comp, arg))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
comp_type.result.compact_representation())
actual_result = self.run_sync(result.compute())
if (all_isinstance([actual_result, expected_result], list)
or all_isinstance([actual_result, expected_result],
tf.data.Dataset)):
for actual_element, expected_element in zip(
actual_result, expected_result):
self.assertEqual(actual_element, expected_element)
else:
self.assertEqual(actual_result, expected_result)
def test_returns_value_with_intrinsic_def_federated_value_at_server_and_tuple(
self):
executor = create_test_executor(number_of_clients=3)
arg, arg_type = executor_test_utils.create_dummy_computation_tuple()
intrinsic_def = intrinsic_defs.FEDERATED_VALUE_AT_SERVER
comp_type = computation_types.FunctionType(
arg_type, type_factory.at_server(arg_type))
comp = pb.Computation(
type=type_serialization.serialize_type(comp_type),
intrinsic=pb.Intrinsic(uri=intrinsic_def.uri))
comp = self.run_sync(executor.create_value(comp, comp_type))
arg = self.run_sync(executor.create_value(arg, arg_type))
result = self.run_sync(executor.create_call(comp, arg))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
comp_type.result.compact_representation())
actual_result = self.run_sync(result.compute())
expected_result = [10.0] * 2
for actual_element, expected_element in zip(actual_result,
expected_result):
self.assertEqual(actual_element, expected_element)
def test_returns_value_with_intrinsic_def_federated_eval_at_clients_and_random(
self):
executor = create_test_executor(number_of_clients=3)
comp, comp_type = executor_test_utils.create_dummy_intrinsic_def_federated_eval_at_clients(
)
arg, arg_type = executor_test_utils.create_dummy_computation_tensorflow_random(
)
comp = self.run_sync(executor.create_value(comp, comp_type))
arg = self.run_sync(executor.create_value(arg, arg_type))
result = self.run_sync(executor.create_call(comp, arg))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
comp_type.result.compact_representation())
actual_result = self.run_sync(result.compute())
unique_results = set([x.numpy() for x in actual_result])
if len(actual_result) != len(unique_results):
self.fail(
'Expected the result to contain different random numbers, found {}.'
.format(actual_result))
# pyformat: disable
@parameterized.named_parameters([
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
])
# pyformat: enable
def test_returns_value_with_comp_only(self, comp, comp_type):
executor = create_test_executor()
comp = self.run_sync(executor.create_value(comp, comp_type))
result = self.run_sync(executor.create_call(comp))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
comp_type.result.compact_representation())
actual_result = self.run_sync(result.compute())
expected_result = []
self.assertCountEqual(actual_result, expected_result)
def test_raises_type_error_with_unembedded_comp(self):
executor = create_test_executor()
comp, _ = executor_test_utils.create_dummy_computation_tensorflow_identity(
)
arg, arg_type = executor_test_utils.create_dummy_value_unplaced()
arg = self.run_sync(executor.create_value(arg, arg_type))
with self.assertRaises(TypeError):
self.run_sync(executor.create_call(comp, arg))
def test_raises_type_error_with_unembedded_arg(self):
executor = create_test_executor()
comp, comp_type = executor_test_utils.create_dummy_computation_tensorflow_identity(
)
arg, _ = executor_test_utils.create_dummy_value_unplaced()
comp = self.run_sync(executor.create_value(comp, comp_type))
with self.assertRaises(TypeError):
self.run_sync(executor.create_call(comp, arg))
# pyformat: disable
@parameterized.named_parameters([
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic()),
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_identity()),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_identity()),
] + get_named_parameters_for_supported_intrinsics())
# pyformat: enable
def test_raises_type_error_with_comp_and_bad_arg(self, comp, comp_type):
executor = create_test_executor()
bad_arg = 'string'
bad_arg_type = computation_types.TensorType(tf.string)
comp = self.run_sync(executor.create_value(comp, comp_type))
arg = self.run_sync(executor.create_value(bad_arg, bad_arg_type))
with self.assertRaises(TypeError):
self.run_sync(executor.create_call(comp, arg))
# pyformat: disable
@parameterized.named_parameters([
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_empty()),
('computation_placement',
*executor_test_utils.create_dummy_computation_placement()),
('computation_tuple',
*executor_test_utils.create_dummy_computation_tuple()),
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
('unplaced_type', *executor_test_utils.create_dummy_value_unplaced()),
])
# pyformat: enable
def test_raises_value_error_with_comp(self, comp, comp_type):
executor = create_test_executor()
comp = self.run_sync(executor.create_value(comp, comp_type))
with self.assertRaises(ValueError):
self.run_sync(executor.create_call(comp))
def test_raises_not_implemented_error_with_intrinsic_def_federated_secure_sum(
self):
executor = create_test_executor()
comp, comp_type = executor_test_utils.create_dummy_intrinsic_def_federated_secure_sum(
)
arg_1, arg_1_type = executor_test_utils.create_dummy_value_at_clients()
arg_2, arg_2_type = executor_test_utils.create_dummy_value_unplaced()
comp = self.run_sync(executor.create_value(comp, comp_type))
arg_1 = self.run_sync(executor.create_value(arg_1, arg_1_type))
arg_2 = self.run_sync(executor.create_value(arg_2, arg_2_type))
args = self.run_sync(executor.create_tuple([arg_1, arg_2]))
with self.assertRaises(NotImplementedError):
self.run_sync(executor.create_call(comp, args))
def test_raises_not_implemented_error_with_unimplemented_intrinsic(self):
executor = create_test_executor()
dummy_intrinsic = intrinsic_defs.IntrinsicDef(
'DUMMY_INTRINSIC', 'dummy_intrinsic',
computation_types.AbstractType('T'))
comp = pb.Computation(intrinsic=pb.Intrinsic(uri='dummy_intrinsic'),
type=type_serialization.serialize_type(tf.int32))
comp = self.run_sync(executor.create_value(comp))
with self.assertRaises(NotImplementedError):
self.run_sync(executor.create_call(comp))
class FederatingExecutorCreateValueTest(executor_test_utils.AsyncTestCase,
parameterized.TestCase):
# pyformat: disable
@parameterized.named_parameters([
('placement_literal',
*executor_test_utils.create_dummy_placement_literal()),
('computation_call',
*executor_test_utils.create_dummy_computation_call()),
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic()),
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_empty()),
('computation_placement',
*executor_test_utils.create_dummy_computation_placement()),
('computation_selection',
*executor_test_utils.create_dummy_computation_selection()),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
('computation_tuple',
*executor_test_utils.create_dummy_computation_tuple()),
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
('unplaced_type', *executor_test_utils.create_dummy_value_unplaced()),
] + get_named_parameters_for_supported_intrinsics())
# pyformat: enable
def test_returns_value_with_value_and_type(self, value, type_signature):
executor = create_test_executor()
result = self.run_sync(executor.create_value(value, type_signature))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
# pyformat: disable
@parameterized.named_parameters([
('placement_literal',
*executor_test_utils.create_dummy_placement_literal()),
('computation_call',
*executor_test_utils.create_dummy_computation_call()),
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic()),
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_empty()),
('computation_placement',
*executor_test_utils.create_dummy_computation_placement()),
('computation_selection',
*executor_test_utils.create_dummy_computation_selection()),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
('computation_tuple',
*executor_test_utils.create_dummy_computation_tuple()),
])
# pyformat: enable
def test_returns_value_with_value_only(self, value, type_signature):
executor = create_test_executor()
result = self.run_sync(executor.create_value(value))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
# pyformat: disable
@parameterized.named_parameters([
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic()),
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_empty()),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
])
# pyformat: enable
def test_returns_value_with_computation_impl(self, proto, type_signature):
executor = create_test_executor()
value = computation_impl.ComputationImpl(
proto, context_stack_impl.context_stack)
result = self.run_sync(executor.create_value(value, type_signature))
self.assertIsInstance(result,
federating_executor.FederatingExecutorValue)
self.assertEqual(result.type_signature.compact_representation(),
type_signature.compact_representation())
# pyformat: disable
@parameterized.named_parameters([
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
('unplaced_type', *executor_test_utils.create_dummy_value_unplaced()),
] + get_named_parameters_for_supported_intrinsics())
# pyformat: enable
def test_raises_type_error_with_value_only(self, value, type_signature):
del type_signature # Unused.
executor = create_test_executor()
with self.assertRaises(TypeError):
self.run_sync(executor.create_value(value))
# pyformat: disable
@parameterized.named_parameters([
('placement_literal',
*executor_test_utils.create_dummy_placement_literal()),
('computation_placement',
*executor_test_utils.create_dummy_computation_placement()),
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
('unplaced_type', *executor_test_utils.create_dummy_value_unplaced()),
] + get_named_parameters_for_supported_intrinsics())
# pyformat: enable
def test_raises_type_error_with_value_and_bad_type(self, value,
type_signature):
del type_signature # Unused.
executor = create_test_executor()
bad_type_signature = computation_types.TensorType(tf.string)
with self.assertRaises(TypeError):
self.run_sync(executor.create_value(value, bad_type_signature))
# pyformat: disable
@parameterized.named_parameters([
('computation_call',
*executor_test_utils.create_dummy_computation_call()),
('computation_intrinsic',
*executor_test_utils.create_dummy_computation_intrinsic()),
('computation_lambda',
*executor_test_utils.create_dummy_computation_lambda_empty()),
('computation_selection',
*executor_test_utils.create_dummy_computation_selection()),
('computation_tensorflow',
*executor_test_utils.create_dummy_computation_tensorflow_empty()),
('computation_tuple',
*executor_test_utils.create_dummy_computation_tuple()),
])
# pyformat: enable
def test_raises_type_error_with_value_and_bad_type_skipped(
self, value, type_signature):
del type_signature # Unused.
self.skipTest(
'TODO(b/152449402): `FederatingExecutor.create_value` method should '
'fail if it is passed a computation and an incompatible type.')
executor = create_test_executor()
bad_type_signature = computation_types.TensorType(tf.string)
with self.assertRaises(TypeError):
self.run_sync(executor.create_value(value, bad_type_signature))
# pyformat: disable
@parameterized.named_parameters([
('computation_reference',
*executor_test_utils.create_dummy_computation_reference()),
('function_type', lambda: 10, type_factory.unary_op(tf.int32)),
])
# pyformat: enable
def test_raises_value_error_with_value(self, value, type_signature):
executor = create_test_executor()
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
def test_raises_value_error_with_unrecognized_computation_intrinsic(self):
executor = create_test_executor()
# A `ValueError` will be raised because `create_value` can not recognize the
# following intrinsic, because it has not been added to the intrinsic
# registry.
value = pb.Computation(
type=type_serialization.serialize_type(tf.int32),
intrinsic=pb.Intrinsic(uri='unregistered_intrinsic'))
type_signature = computation_types.TensorType(tf.int32)
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
def test_raises_value_error_with_unrecognized_computation_selection(self):
executor = create_test_executor()
source, _ = executor_test_utils.create_dummy_computation_tuple()
type_signature = computation_types.NamedTupleType([])
# A `ValueError` will be raised because `create_value` can not handle the
# following `pb.Selection`, because does not set either a name or an index
# field.
value = pb.Computation(
type=type_serialization.serialize_type(type_signature),
selection=pb.Selection(source=source))
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
# pyformat: disable
@parameterized.named_parameters([
('intrinsic_def_federated_broadcast',
*executor_test_utils.create_dummy_intrinsic_def_federated_broadcast()
),
('intrinsic_def_federated_eval_at_clients', *executor_test_utils.
create_dummy_intrinsic_def_federated_eval_at_clients()),
('intrinsic_def_federated_map',
*executor_test_utils.create_dummy_intrinsic_def_federated_map()),
('intrinsic_def_federated_map_all_equal', *executor_test_utils.
create_dummy_intrinsic_def_federated_map_all_equal()),
('intrinsic_def_federated_value_at_clients', *executor_test_utils.
create_dummy_intrinsic_def_federated_value_at_clients()),
('federated_type_at_clients_all_equal',
*executor_test_utils.create_dummy_value_at_clients_all_equal()),
('federated_type_at_clients',
*executor_test_utils.create_dummy_value_at_clients())
])
# pyformat: enable
def test_raises_value_error_with_no_target_executor_clients(
self, value, type_signature):
executor = federating_executor.FederatingExecutor({
placement_literals.SERVER:
eager_tf_executor.EagerTFExecutor(),
None:
eager_tf_executor.EagerTFExecutor()
})
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
# pyformat: disable
@parameterized.named_parameters([
('intrinsic_def_federated_aggregate',
*executor_test_utils.create_dummy_intrinsic_def_federated_aggregate()
),
('intrinsic_def_federated_apply',
*executor_test_utils.create_dummy_intrinsic_def_federated_apply()),
('intrinsic_def_federated_collect',
*executor_test_utils.create_dummy_intrinsic_def_federated_collect()),
('intrinsic_def_federated_eval_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_eval_at_server()),
('intrinsic_def_federated_mean',
*executor_test_utils.create_dummy_intrinsic_def_federated_mean()),
('intrinsic_def_federated_sum',
*executor_test_utils.create_dummy_intrinsic_def_federated_sum()),
('intrinsic_def_federated_reduce',
*executor_test_utils.create_dummy_intrinsic_def_federated_reduce()),
('intrinsic_def_federated_value_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_value_at_server()),
('intrinsic_def_federated_weighted_mean', *executor_test_utils.
create_dummy_intrinsic_def_federated_weighted_mean()),
('intrinsic_def_federated_zip_at_server', *executor_test_utils.
create_dummy_intrinsic_def_federated_zip_at_server()),
('federated_type_at_server',
*executor_test_utils.create_dummy_value_at_server()),
])
# pyformat: enable
def test_raises_value_error_with_no_target_executor_server(
self, value, type_signature):
executor = federating_executor.FederatingExecutor({
placement_literals.CLIENTS:
eager_tf_executor.EagerTFExecutor(),
None:
eager_tf_executor.EagerTFExecutor()
})
value, type_signature = executor_test_utils.create_dummy_value_at_server(
)
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
def test_raises_value_error_with_no_target_executor_unplaced(self):
executor = federating_executor.FederatingExecutor({
placement_literals.SERVER:
eager_tf_executor.EagerTFExecutor(),
placement_literals.CLIENTS:
eager_tf_executor.EagerTFExecutor(),
})
value, type_signature = executor_test_utils.create_dummy_value_unplaced(
)
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
def test_raises_value_error_with_unexpected_federated_type_at_clients(
self):
executor = create_test_executor()
value = [10, 20]
type_signature = type_factory.at_clients(tf.int32)
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
def test_raises_value_error_with_unexpected_federated_type_at_clients_all_equal(
self):
executor = create_test_executor()
value = [10] * 3
type_signature = type_factory.at_clients(tf.int32, all_equal=True)
with self.assertRaises(ValueError):
self.run_sync(executor.create_value(value, type_signature))
