def test_call_fit_none_mini_batch_size(sagemaker_session):
ipinsights = IPInsights(base_job_name="ipinsights",
sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel='train')
ipinsights.fit(data)
def test_model_image(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel='train')
ipinsights.fit(data, MINI_BATCH_SIZE)
model = ipinsights.create_model()
assert model.image == registry(REGION, "ipinsights") + '/ipinsights:1'
def test_model_image(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
ipinsights.fit(data, MINI_BATCH_SIZE)
model = ipinsights.create_model()
assert image_uris.retrieve("ipinsights", REGION) == model.image_uri
def test_predictor_type(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel='train')
ipinsights.fit(data, MINI_BATCH_SIZE)
model = ipinsights.create_model()
predictor = model.deploy(1, TRAIN_INSTANCE_TYPE)
assert isinstance(predictor, IPInsightsPredictor)
def test_prepare_for_training_wrong_value_upper_mini_batch_size(
sagemaker_session):
ipinsights = IPInsights(base_job_name="ipinsights",
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):
ipinsights._prepare_for_training(data, 500001)
def test_call_fit(base_fit, sagemaker_session):
ipinsights = IPInsights(base_job_name="ipinsights",
sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel='train')
ipinsights.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_init_required_named(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert ipinsights.role == COMMON_TRAIN_ARGS["role"]
assert ipinsights.instance_count == INSTANCE_COUNT
assert ipinsights.instance_type == COMMON_TRAIN_ARGS["instance_type"]
assert ipinsights.num_entity_vectors == NUM_ENTITY_VECTORS
assert ipinsights.vector_dim == VECTOR_DIM
def test_all_hyperparameters(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
batch_metrics_publish_interval=100,
epochs=10,
learning_rate=0.001,
num_ip_encoder_layers=3,
random_negative_sampling_rate=5,
shuffled_negative_sampling_rate=5,
weight_decay=5.0,
**ALL_REQ_ARGS)
assert ipinsights.hyperparameters() == dict(
num_entity_vectors=str(ALL_REQ_ARGS['num_entity_vectors']),
vector_dim=str(ALL_REQ_ARGS['vector_dim']),
batch_metrics_publish_interval='100',
epochs='10',
learning_rate='0.001',
num_ip_encoder_layers='3',
random_negative_sampling_rate='5',
shuffled_negative_sampling_rate='5',
weight_decay='5.0')
def test_init_required_positional(sagemaker_session):
ipinsights = IPInsights(ROLE,
TRAIN_INSTANCE_COUNT,
TRAIN_INSTANCE_TYPE,
NUM_ENTITY_VECTORS,
VECTOR_DIM,
sagemaker_session=sagemaker_session)
assert ipinsights.role == ROLE
assert ipinsights.train_instance_count == TRAIN_INSTANCE_COUNT
assert ipinsights.train_instance_type == TRAIN_INSTANCE_TYPE
assert ipinsights.num_entity_vectors == NUM_ENTITY_VECTORS
assert ipinsights.vector_dim == VECTOR_DIM
def test_all_hyperparameters(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
batch_metrics_publish_interval=100,
epochs=10,
learning_rate=0.001,
num_ip_encoder_layers=3,
random_negative_sampling_rate=5,
shuffled_negative_sampling_rate=5,
weight_decay=5.0,
**ALL_REQ_ARGS)
assert ipinsights.hyperparameters() == dict(
num_entity_vectors=str(ALL_REQ_ARGS["num_entity_vectors"]),
vector_dim=str(ALL_REQ_ARGS["vector_dim"]),
batch_metrics_publish_interval="100",
epochs="10",
learning_rate="0.001",
num_ip_encoder_layers="3",
random_negative_sampling_rate="5",
shuffled_negative_sampling_rate="5",
weight_decay="5.0",
)
def test_predictor_custom_serialization(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet(
"s3://{}/{}".format(BUCKET_NAME, PREFIX),
num_records=1,
feature_dim=FEATURE_DIM,
channel="train",
)
ipinsights.fit(data, MINI_BATCH_SIZE)
model = ipinsights.create_model()
custom_serializer = Mock()
custom_deserializer = Mock()
predictor = model.deploy(
1,
INSTANCE_TYPE,
serializer=custom_serializer,
deserializer=custom_deserializer,
)
assert isinstance(predictor, IPInsightsPredictor)
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})
IPInsights(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
IPInsights(sagemaker_session=sagemaker_session, **test_params)
def test_image(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert image_uris.retrieve("ipinsights", REGION) == ipinsights.training_image_uri()
def test_image(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
assert ipinsights.train_image() == registry(REGION,
"ipinsights") + '/ipinsights:1'