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_model_image(sagemaker_session):
knn = KNN(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
knn.fit(data, MINI_BATCH_SIZE)
model = knn.create_model()
assert model.image == registry(REGION, "knn") + '/knn:1'
def __init__(self, model_data, role, sagemaker_session=None):
sagemaker_session = sagemaker_session or Session()
repo = '{}:{}'.format(LinearLearner.repo_name, LinearLearner.repo_version)
image = '{}/{}'.format(registry(sagemaker_session.boto_session.region_name), repo)
super(LinearLearnerModel, self).__init__(model_data, image, role,
predictor_cls=LinearLearnerPredictor,
sagemaker_session=sagemaker_session)
def test_model_image(sagemaker_session):
randomcutforest = RandomCutForest(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet("s3://{}/{}".format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
randomcutforest.fit(data, MINI_BATCH_SIZE)
model = randomcutforest.create_model()
assert model.image == registry(REGION, "randomcutforest") + '/randomcutforest:1'
def test_model_image(sagemaker_session):
lr = LinearLearner(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
lr.fit(data)
model = lr.create_model()
assert model.image == registry(REGION, 'linear-learner') + '/linear-learner:1'
def test_model_image(sagemaker_session):
lda = LDA(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
lda.fit(data, MINI_BATCH_SZIE)
model = lda.create_model()
assert model.image == registry(REGION, 'lda') + '/lda:1'
def sagemaker_session():
boto_mock = Mock(name='boto_session', region_name=REGION)
sms = Mock(name='sagemaker_session', boto_session=boto_mock,
region_name=REGION, config=None, local_mode=False)
sms.boto_region_name = REGION
sms.default_bucket = Mock(name='default_bucket', return_value=BUCKET_NAME)
returned_job_description = {'AlgorithmSpecification': {'TrainingInputMode': 'File',
'TrainingImage': registry("us-west-2") + "/pca:1"},
'ModelArtifacts': {'S3ModelArtifacts': "s3://some-bucket/model.tar.gz"},
'HyperParameters':
{'sagemaker_submit_directory': '"s3://some/sourcedir.tar.gz"',
'checkpoint_path': '"s3://other/1508872349"',
'sagemaker_program': '"iris-dnn-classifier.py"',
'sagemaker_enable_cloudwatch_metrics': 'false',
'sagemaker_container_log_level': '"logging.INFO"',
'sagemaker_job_name': '"neo"',
'training_steps': '100'},
'RoleArn': 'arn:aws:iam::366:role/IMRole',
'ResourceConfig':
{'VolumeSizeInGB': 30,
'InstanceCount': 1,
'InstanceType': 'ml.c4.xlarge'},
'StoppingCondition': {'MaxRuntimeInSeconds': 24 * 60 * 60},
'TrainingJobName': 'neo',
'TrainingJobStatus': 'Completed',
'OutputDataConfig': {'KmsKeyId': '',
'S3OutputPath': 's3://place/output/neo'},
'TrainingJobOutput': {'S3TrainingJobOutput': 's3://here/output.tar.gz'}}
sms.sagemaker_client.describe_training_job = Mock(name='describe_training_job',
return_value=returned_job_description)
return sms
def test_model_image(sagemaker_session):
fm = FactorizationMachines(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
data = RecordSet('s3://{}/{}'.format(BUCKET_NAME, PREFIX), num_records=1, feature_dim=FEATURE_DIM, channel='train')
fm.fit(data, MINI_BATCH_SIZE)
model = fm.create_model()
assert model.image == registry(REGION, 'factorization-machines') + '/factorization-machines:1'
def __init__(self, model_data, role, sagemaker_session=None):
sagemaker_session = sagemaker_session or Session()
repo = '{}:{}'.format(RandomCutForest.repo_name, RandomCutForest.repo_version)
image = '{}/{}'.format(registry(sagemaker_session.boto_session.region_name,
RandomCutForest.repo_name), repo)
super(RandomCutForestModel, self).__init__(model_data, image, role,
predictor_cls=RandomCutForestPredictor,
sagemaker_session=sagemaker_session)
def __init__(self, model_data, role, sagemaker_session=None):
sagemaker_session = sagemaker_session or Session()
repo = '{}:{}'.format(FactorizationMachines.repo_name, FactorizationMachines.repo_version)
image = '{}/{}'.format(registry(sagemaker_session.boto_session.region_name), repo)
super(FactorizationMachinesModel, self).__init__(model_data,
image,
role,
predictor_cls=FactorizationMachinesPredictor,
sagemaker_session=sagemaker_session)
def test_async_byo_estimator(sagemaker_session, region):
image_name = registry(region) + "/factorization-machines:1"
endpoint_name = name_from_base('byo')
training_job_name = ""
with timeout(minutes=5):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
# take 100 examples for faster execution
vectors = np.array([t.tolist() for t in train_set[0][:100]]).astype('float32')
labels = np.where(np.array([t.tolist() for t in train_set[1][:100]]) == 0, 1.0, 0.0).astype('float32')
buf = io.BytesIO()
write_numpy_to_dense_tensor(buf, vectors, labels)
buf.seek(0)
bucket = sagemaker_session.default_bucket()
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
boto3.resource('s3').Bucket(bucket).Object(os.path.join(prefix, 'train', key)).upload_fileobj(buf)
s3_train_data = 's3://{}/{}/train/{}'.format(bucket, prefix, key)
estimator = Estimator(image_name=image_name,
role='SageMakerRole', train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session, base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data}, wait=False)
training_job_name = estimator.latest_training_job.name
with timeout_and_delete_endpoint_by_name(endpoint_name, sagemaker_session):
estimator = Estimator.attach(training_job_name=training_job_name, sagemaker_session=sagemaker_session)
model = estimator.create_model()
predictor = model.deploy(1, 'ml.m4.xlarge', endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
assert estimator.train_image() == image_name
def test_byo_estimator(sagemaker_session, region):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(region) + "/factorization-machines:1"
training_data_path = os.path.join(DATA_DIR, 'dummy_tensor')
with timeout(minutes=TRAINING_DEFAULT_TIMEOUT_MINUTES):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
s3_train_data = sagemaker_session.upload_data(path=training_data_path,
key_prefix=os.path.join(prefix, 'train', key))
estimator = Estimator(image_name=image_name,
role='SageMakerRole', train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session, base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data})
endpoint_name = name_from_base('byo')
with timeout_and_delete_endpoint_by_name(endpoint_name, sagemaker_session):
model = estimator.create_model()
predictor = model.deploy(1, 'ml.m4.xlarge', endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
def test_async_byo_estimator(sagemaker_session, region):
image_name = registry(region) + "/factorization-machines:1"
endpoint_name = name_from_base('byo')
training_data_path = os.path.join(DATA_DIR, 'dummy_tensor')
training_job_name = ""
with timeout(minutes=5):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
s3_train_data = sagemaker_session.upload_data(path=training_data_path,
key_prefix=os.path.join(prefix, 'train', key))
estimator = Estimator(image_name=image_name,
role='SageMakerRole', train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session, base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data}, wait=False)
training_job_name = estimator.latest_training_job.name
with timeout_and_delete_endpoint_by_name(endpoint_name, sagemaker_session):
estimator = Estimator.attach(training_job_name=training_job_name, sagemaker_session=sagemaker_session)
model = estimator.create_model()
predictor = model.deploy(1, 'ml.m4.xlarge', endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
assert estimator.train_image() == image_name
def test_image(sagemaker_session):
randomcutforest = RandomCutForest(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert randomcutforest.train_image() == registry(REGION, "randomcutforest") + '/randomcutforest:1'
def test_image(sagemaker_session):
ntm = NTM(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert ntm.train_image() == registry(REGION, "ntm") + '/ntm:1'
def test_tuning_byo_estimator(sagemaker_session):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(sagemaker_session.boto_session.region_name) + "/factorization-machines:1"
training_data_path = os.path.join(DATA_DIR, "dummy_tensor")
with timeout(minutes=TUNING_DEFAULT_TIMEOUT_MINUTES):
data_path = os.path.join(DATA_DIR, "one_p_mnist", "mnist.pkl.gz")
pickle_args = {} if sys.version_info.major == 2 else {"encoding": "latin1"}
with gzip.open(data_path, "rb") as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = "test_byo_estimator"
key = "recordio-pb-data"
s3_train_data = sagemaker_session.upload_data(
path=training_data_path, key_prefix=os.path.join(prefix, "train", key)
)
estimator = Estimator(
image_name=image_name,
role="SageMakerRole",
train_instance_count=1,
train_instance_type="ml.c4.xlarge",
sagemaker_session=sagemaker_session,
)
estimator.set_hyperparameters(
num_factors=10, feature_dim=784, mini_batch_size=100, predictor_type="binary_classifier"
)
hyperparameter_ranges = {"mini_batch_size": IntegerParameter(100, 200)}
tuner = HyperparameterTuner(
estimator=estimator,
objective_metric_name="test:binary_classification_accuracy",
hyperparameter_ranges=hyperparameter_ranges,
max_jobs=2,
max_parallel_jobs=2,
)
tuner.fit(
{"train": s3_train_data, "test": s3_train_data},
include_cls_metadata=False,
job_name=unique_name_from_base("byo", 32),
)
print("Started hyperparameter tuning job with name:" + tuner.latest_tuning_job.name)
time.sleep(15)
tuner.wait()
best_training_job = tuner.best_training_job()
with timeout_and_delete_endpoint_by_name(best_training_job, sagemaker_session):
predictor = tuner.deploy(1, "ml.m4.xlarge", endpoint_name=best_training_job)
predictor.serializer = _fm_serializer
predictor.content_type = "application/json"
predictor.deserializer = json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result["predictions"]) == 10
for prediction in result["predictions"]:
assert prediction["score"] is not None
def test_image(sagemaker_session):
pca = PCA(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert pca.train_image() == registry(REGION, 'pca') + '/pca:1'
def test_regitry_throws_error_if_mapping_does_not_exist_for_default_algorithm(
):
with pytest.raises(ValueError) as error:
registry("broken_region_name")
assert "Algorithm (None) is unsupported for region (broken_region_name)." in str(
error)
def test_image(sagemaker_session):
lda = LDA(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert lda.train_image() == registry(REGION, 'lda') + '/lda:1'
def test_image(sagemaker_session):
fm = FactorizationMachines(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
assert fm.train_image() == registry(REGION) + "/factorization-machines:1"
def test_byo_estimator():
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(REGION) + "/factorization-machines:1"
with timeout(minutes=15):
sagemaker_session = sagemaker.Session(boto_session=boto3.Session(
region_name=REGION))
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {
'encoding': 'latin1'
}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
# take 100 examples for faster execution
vectors = np.array([t.tolist()
for t in train_set[0][:100]]).astype('float32')
labels = np.where(
np.array([t.tolist() for t in train_set[1][:100]]) == 0, 1.0,
0.0).astype('float32')
buf = io.BytesIO()
write_numpy_to_dense_tensor(buf, vectors, labels)
buf.seek(0)
bucket = sagemaker_session.default_bucket()
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
boto3.resource('s3').Bucket(bucket).Object(
os.path.join(prefix, 'train', key)).upload_fileobj(buf)
s3_train_data = 's3://{}/{}/train/{}'.format(bucket, prefix, key)
estimator = Estimator(image_name=image_name,
role='SageMakerRole',
train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session,
base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data})
endpoint_name = name_from_base('byo')
def fm_serializer(data):
js = {'instances': []}
for row in data:
js['instances'].append({'features': row.tolist()})
return json.dumps(js)
with timeout_and_delete_endpoint_by_name(endpoint_name,
sagemaker_session,
minutes=20):
model = estimator.create_model()
predictor = model.deploy(1,
'ml.m4.xlarge',
endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
def test_tuning_byo_estimator(sagemaker_session):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(sagemaker_session.boto_session.region_name
) + '/factorization-machines:1'
training_data_path = os.path.join(DATA_DIR, 'dummy_tensor')
with timeout(minutes=TUNING_DEFAULT_TIMEOUT_MINUTES):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {
'encoding': 'latin1'
}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
s3_train_data = sagemaker_session.upload_data(path=training_data_path,
key_prefix=os.path.join(
prefix, 'train',
key))
estimator = Estimator(image_name=image_name,
role='SageMakerRole',
train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session,
base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
hyperparameter_ranges = {'mini_batch_size': IntegerParameter(100, 200)}
tuner = HyperparameterTuner(
estimator=estimator,
base_tuning_job_name='byo',
objective_metric_name='test:binary_classification_accuracy',
hyperparameter_ranges=hyperparameter_ranges,
max_jobs=2,
max_parallel_jobs=2)
tuner.fit({
'train': s3_train_data,
'test': s3_train_data
},
include_cls_metadata=False)
print('Started hyperparameter tuning job with name:' +
tuner.latest_tuning_job.name)
time.sleep(15)
tuner.wait()
best_training_job = tuner.best_training_job()
with timeout_and_delete_endpoint_by_name(best_training_job,
sagemaker_session):
predictor = tuner.deploy(1,
'ml.m4.xlarge',
endpoint_name=best_training_job)
predictor.serializer = _fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
def test_learning_rate_validation_fail_value_upper(sagemaker_session):
with pytest.raises(ValueError):
NTM(learning_rate=2,
sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
PREFIX = "prefix"
BASE_TRAIN_CALL = {
'hyperparameters': {},
'image':
registry(REGION, "ntm") + '/ntm:1',
'input_config': [{
'DataSource': {
'S3DataSource': {
'S3DataDistributionType': 'ShardedByS3Key',
'S3DataType': 'ManifestFile',
'S3Uri': 's3://{}/{}'.format(BUCKET_NAME, PREFIX)
}
},
'ChannelName': 'train'
}],
'input_mode':
'File',
'output_config': {
'S3OutputPath': 's3://{}/'.format(BUCKET_NAME)
},
def test_image(sagemaker_session):
knn = KNN(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert knn.train_image() == registry(REGION, "knn") + '/knn:1'
def test_image(sagemaker_session):
pca = PCA(num_components=55,
sagemaker_session=sagemaker_session,
**COMMON_ARGS)
assert pca.train_image() == registry('us-west-2') + '/pca:1'
def test_image(sagemaker_session):
randomcutforest = RandomCutForest(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert (
randomcutforest.train_image() == registry(REGION, "randomcutforest") + "/randomcutforest:1"
)
def test_image(sagemaker_session):
lr = LinearLearner(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert lr.train_image() == registry(REGION,
'linear-learner') + '/linear-learner:1'
def test_byo_estimator(sagemaker_session, region):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(region) + "/factorization-machines:1"
training_data_path = os.path.join(DATA_DIR, 'dummy_tensor')
with timeout(minutes=15):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {
'encoding': 'latin1'
}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
s3_train_data = sagemaker_session.upload_data(path=training_data_path,
key_prefix=os.path.join(
prefix, 'train',
key))
estimator = Estimator(image_name=image_name,
role='SageMakerRole',
train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session,
base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data})
endpoint_name = name_from_base('byo')
with timeout_and_delete_endpoint_by_name(endpoint_name, sagemaker_session):
model = estimator.create_model()
predictor = model.deploy(1,
'ml.m4.xlarge',
endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
def test_image(sagemaker_session):
ipinsights = IPInsights(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
assert ipinsights.train_image() == registry(REGION,
"ipinsights") + '/ipinsights:1'
def test_image(sagemaker_session):
lr = LinearLearner(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert lr.train_image() == registry(REGION, 'linear-learner') + '/linear-learner:1'
def test_image(sagemaker_session):
lda = LDA(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert lda.train_image() == registry(REGION, "lda") + "/lda:1"
def test_image(sagemaker_session):
knn = KNN(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert knn.train_image() == registry(REGION, "knn") + '/knn:1'
def test_regitry_throws_error_if_mapping_does_not_exist_for_lda():
with pytest.raises(ValueError) as error:
registry("cn-north-1", "lda")
assert "Algorithm (lda) is unsupported for region (cn-north-1)." in str(
error)
def test_tuning_byo_estimator(sagemaker_session):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(sagemaker_session.boto_session.region_name) + '/factorization-machines:1'
training_data_path = os.path.join(DATA_DIR, 'dummy_tensor')
with timeout(minutes=TUNING_DEFAULT_TIMEOUT_MINUTES):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
s3_train_data = sagemaker_session.upload_data(path=training_data_path,
key_prefix=os.path.join(prefix, 'train', key))
estimator = Estimator(image_name=image_name,
role='SageMakerRole', train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session, base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
hyperparameter_ranges = {'mini_batch_size': IntegerParameter(100, 200)}
tuner = HyperparameterTuner(estimator=estimator, base_tuning_job_name='byo',
objective_metric_name='test:binary_classification_accuracy',
hyperparameter_ranges=hyperparameter_ranges,
max_jobs=2, max_parallel_jobs=2)
tuner.fit({'train': s3_train_data, 'test': s3_train_data}, include_cls_metadata=False)
print('Started hyperparameter tuning job with name:' + tuner.latest_tuning_job.name)
time.sleep(15)
tuner.wait()
best_training_job = tuner.best_training_job()
with timeout_and_delete_endpoint_by_name(best_training_job, sagemaker_session):
predictor = tuner.deploy(1, 'ml.m4.xlarge', endpoint_name=best_training_job)
predictor.serializer = _fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
def test_image(sagemaker_session):
ntm = NTM(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert ntm.train_image() == registry(REGION, "ntm") + '/ntm:1'
def __init__(self, model_data, role, sagemaker_session=None):
sagemaker_session = sagemaker_session or Session()
image = registry(sagemaker_session.boto_session.region_name) + "/" + PCA.repo
super(PCAModel, self).__init__(model_data, image, role, predictor_cls=PCAPredictor,
sagemaker_session=sagemaker_session)
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_byo_estimator(sagemaker_session, region):
"""Use Factorization Machines algorithm as an example here.
First we need to prepare data for training. We take standard data set, convert it to the
format that the algorithm can process and upload it to S3.
Then we create the Estimator and set hyperparamets as required by the algorithm.
Next, we can call fit() with path to the S3.
Later the trained model is deployed and prediction is called against the endpoint.
Default predictor is updated with json serializer and deserializer.
"""
image_name = registry(region) + "/factorization-machines:1"
with timeout(minutes=15):
data_path = os.path.join(DATA_DIR, 'one_p_mnist', 'mnist.pkl.gz')
pickle_args = {} if sys.version_info.major == 2 else {'encoding': 'latin1'}
with gzip.open(data_path, 'rb') as f:
train_set, _, _ = pickle.load(f, **pickle_args)
# take 100 examples for faster execution
vectors = np.array([t.tolist() for t in train_set[0][:100]]).astype('float32')
labels = np.where(np.array([t.tolist() for t in train_set[1][:100]]) == 0, 1.0, 0.0).astype('float32')
buf = io.BytesIO()
write_numpy_to_dense_tensor(buf, vectors, labels)
buf.seek(0)
bucket = sagemaker_session.default_bucket()
prefix = 'test_byo_estimator'
key = 'recordio-pb-data'
boto3.resource('s3').Bucket(bucket).Object(os.path.join(prefix, 'train', key)).upload_fileobj(buf)
s3_train_data = 's3://{}/{}/train/{}'.format(bucket, prefix, key)
estimator = Estimator(image_name=image_name,
role='SageMakerRole', train_instance_count=1,
train_instance_type='ml.c4.xlarge',
sagemaker_session=sagemaker_session, base_job_name='test-byo')
estimator.set_hyperparameters(num_factors=10,
feature_dim=784,
mini_batch_size=100,
predictor_type='binary_classifier')
# training labels must be 'float32'
estimator.fit({'train': s3_train_data})
endpoint_name = name_from_base('byo')
with timeout_and_delete_endpoint_by_name(endpoint_name, sagemaker_session):
model = estimator.create_model()
predictor = model.deploy(1, 'ml.m4.xlarge', endpoint_name=endpoint_name)
predictor.serializer = fm_serializer
predictor.content_type = 'application/json'
predictor.deserializer = sagemaker.predictor.json_deserializer
result = predictor.predict(train_set[0][:10])
assert len(result['predictions']) == 10
for prediction in result['predictions']:
assert prediction['score'] is not None
def test_image(sagemaker_session):
object2vec = Object2Vec(sagemaker_session=sagemaker_session,
**ALL_REQ_ARGS)
assert object2vec.train_image() == registry(REGION,
"object2vec") + "/object2vec:1"
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):
fm = FactorizationMachines(sagemaker_session=sagemaker_session, **ALL_REQ_ARGS)
assert fm.train_image() == registry(REGION) + '/factorization-machines:1'
