def test_model_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
assert model.image == registry(REGION, 'kmeans') + '/kmeans:1'
def test_all_hyperparameters(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session,
init_method='random',
max_iterations=3,
tol=0.5,
num_trials=5,
local_init_method='kmeans++',
half_life_time_size=0,
epochs=10,
center_factor=2,
eval_metrics=['msd', 'ssd'],
**ALL_REQ_ARGS)
assert kmeans.hyperparameters() == dict(
k=str(ALL_REQ_ARGS['k']),
init_method='random',
local_lloyd_max_iter='3',
local_lloyd_tol='0.5',
local_lloyd_num_trials='5',
local_lloyd_init_method='kmeans++',
half_life_time_size='0',
epochs='10',
extra_center_factor='2',
eval_metrics='[\'msd\', \'ssd\']',
force_dense='True',
)
def test_model_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
assert model.image == registry(REGION, 'kmeans') + '/kmeans:1'
def test_prepare_for_training_wrong_value_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name='kmeans', sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM,
channel='train')
with pytest.raises(ValueError):
kmeans._prepare_for_training(data, 0)
def test_call_fit_wrong_value_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name='kmeans', sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM,
channel='train')
with pytest.raises(ValueError):
kmeans.fit(data, 0)
def test_all_hyperparameters(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session,
init_method="random",
max_iterations=3,
tol=0.5,
num_trials=5,
local_init_method="kmeans++",
half_life_time_size=0,
epochs=10,
center_factor=2,
eval_metrics=["msd", "ssd"],
**ALL_REQ_ARGS)
assert kmeans.hyperparameters() == dict(
k=str(ALL_REQ_ARGS["k"]),
init_method="random",
local_lloyd_max_iter="3",
local_lloyd_tol="0.5",
local_lloyd_num_trials="5",
local_lloyd_init_method="kmeans++",
half_life_time_size="0",
epochs="10",
extra_center_factor="2",
eval_metrics='["msd", "ssd"]',
force_dense="True",
)
def test_predictor_type(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
predictor = model.deploy(1, TRAIN_INSTANCE_TYPE)
assert isinstance(predictor, KMeansPredictor)
def test_prepare_for_training_no_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name='kmeans', sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM,
channel='train')
kmeans._prepare_for_training(data)
assert kmeans.mini_batch_size == 5000
def test_predictor_type(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
predictor = model.deploy(1, TRAIN_INSTANCE_TYPE)
assert isinstance(predictor, KMeansPredictor)
def test_call_fit(base_fit, sagemaker_session):
kmeans = KMeans(base_job_name='kmeans', sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
base_fit.assert_called_once()
assert len(base_fit.call_args[0]) == 2
assert base_fit.call_args[0][0] == data
assert base_fit.call_args[0][1] == MINI_BATCH_SIZE
def test_call_fit(base_fit, sagemaker_session):
kmeans = KMeans(base_job_name='kmeans', sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
kmeans.fit(data, MINI_BATCH_SIZE)
base_fit.assert_called_once()
assert len(base_fit.call_args[0]) == 2
assert base_fit.call_args[0][0] == data
assert base_fit.call_args[0][1] == MINI_BATCH_SIZE
def test_prepare_for_training_wrong_value_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name="kmeans", sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
with pytest.raises(ValueError):
kmeans._prepare_for_training(data, 0)
def test_model_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
assert image_uris.retrieve("kmeans", REGION) == model.image_uri
def test_prepare_for_training_wrong_type_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name='kmeans',
sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel='train')
with pytest.raises((TypeError, ValueError)):
kmeans._prepare_for_training(data, 'some')
def test_prepare_for_training_no_mini_batch_size(sagemaker_session):
kmeans = KMeans(base_job_name="kmeans", sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
kmeans._prepare_for_training(data)
assert kmeans.mini_batch_size == 5000
def test_init_required_named(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert kmeans.role == COMMON_TRAIN_ARGS['role']
assert kmeans.train_instance_count == TRAIN_INSTANCE_COUNT
assert kmeans.train_instance_type == COMMON_TRAIN_ARGS['train_instance_type']
assert kmeans.k == ALL_REQ_ARGS['k']
def test_init_required_named(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert kmeans.role == COMMON_TRAIN_ARGS["role"]
assert kmeans.instance_count == INSTANCE_COUNT
assert kmeans.instance_type == COMMON_TRAIN_ARGS["instance_type"]
assert kmeans.k == ALL_REQ_ARGS["k"]
def test_init_required_positional(sagemaker_session):
kmeans = KMeans(
ROLE, TRAIN_INSTANCE_COUNT, TRAIN_INSTANCE_TYPE, K, sagemaker_session=sagemaker_session
)
assert kmeans.role == ROLE
assert kmeans.train_instance_count == TRAIN_INSTANCE_COUNT
assert kmeans.train_instance_type == TRAIN_INSTANCE_TYPE
assert kmeans.k == K
def test_all_hyperparameters(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, init_method='random', max_iterations=3, tol=0.5,
num_trials=5, local_init_method='kmeans++', half_life_time_size=0, epochs=10, center_factor=2,
eval_metrics=['msd', 'ssd'], **ALL_REQ_ARGS)
assert kmeans.hyperparameters() == dict(
k=str(ALL_REQ_ARGS['k']),
init_method='random',
local_lloyd_max_iter='3',
local_lloyd_tol='0.5',
local_lloyd_num_trials='5',
local_lloyd_init_method='kmeans++',
half_life_time_size='0',
epochs='10',
extra_center_factor='2',
eval_metrics='[\'msd\', \'ssd\']',
force_dense='True',
)
def test_predictor_custom_serialization(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
kmeans.fit(data, MINI_BATCH_SIZE)
model = kmeans.create_model()
custom_serializer = Mock()
custom_deserializer = Mock()
predictor = model.deploy(
1,
INSTANCE_TYPE,
serializer=custom_serializer,
deserializer=custom_deserializer,
)
assert isinstance(predictor, KMeansPredictor)
assert predictor.serializer is custom_serializer
assert predictor.deserializer is custom_deserializer
def test_optional_hyper_parameters_value(sagemaker_session,
optional_hyper_parameters, value):
with pytest.raises(ValueError):
test_params = ALL_REQ_ARGS.copy()
test_params.update({optional_hyper_parameters: value})
KMeans(sagemaker_session=sagemaker_session, **test_params)
def test_iterable_hyper_parameters_type(sagemaker_session,
iterable_hyper_parameters, value):
with pytest.raises(TypeError):
test_params = ALL_REQ_ARGS.copy()
test_params.update({iterable_hyper_parameters: value})
KMeans(sagemaker_session=sagemaker_session, **test_params)
def test_required_hyper_parameters_value(sagemaker_session,
required_hyper_parameters, value):
with pytest.raises(ValueError):
test_params = ALL_REQ_ARGS.copy()
test_params[required_hyper_parameters] = value
KMeans(sagemaker_session=sagemaker_session, **test_params)
def test_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert kmeans.train_image() == registry(REGION, "kmeans") + "/kmeans:1"
def test_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert image_uris.retrieve("kmeans", REGION) == kmeans.training_image_uri()
def test_fit_1p(sagemaker_session):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
# Load the data into memory as numpy arrays
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
kmeans = KMeans(role='SageMakerRole',
train_instance_count=1,
train_instance_type='ml.c4.xlarge',
k=10,
sagemaker_session=sagemaker_session,
base_job_name='tk',
output_path='s3://{}/'.format(
sagemaker_session.default_bucket()))
# set kmeans specific hp
kmeans.init_method = 'random'
kmeans.max_iterators = 1
kmeans.tol = 1
kmeans.num_trials = 1
kmeans.local_init_method = 'kmeans++'
kmeans.half_life_time_size = 1
kmeans.epochs = 1
records = kmeans.record_set(train_set[0][:100])
test_records = kmeans.record_set(train_set[0][:100], channel='test')
# specify which hp you want to optimize over
hyperparameter_ranges = {
'extra_center_factor': IntegerParameter(1, 10),
'mini_batch_size': IntegerParameter(10, 100),
'epochs': IntegerParameter(1, 2),
'init_method': CategoricalParameter(['kmeans++', 'random'])
}
objective_metric_name = 'test:msd'
tuner = HyperparameterTuner(estimator=kmeans,
objective_metric_name=objective_metric_name,
hyperparameter_ranges=hyperparameter_ranges,
objective_type='Minimize',
max_jobs=2,
max_parallel_jobs=2)
tuner.fit([records, test_records])
print('Started HPO job with name:' + tuner.latest_tuning_job.name)
def test_image(sagemaker_session):
kmeans = KMeans(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert kmeans.train_image() == registry(REGION, 'kmeans') + '/kmeans:1'
