def main():
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
# For local training a dummy role will be sufficient
role = 'arn:aws:iam::111111111111:role/service-role/AmazonSageMaker-ExecutionRole-20200101T000001'
print('Deploying local mode endpoint')
print(
'Note: if launching for the first time in local mode, container image download might take a few minutes to complete.'
)
pytorch_model = PyTorchModel(model_data='./model/model.tar.gz',
role=role,
framework_version="1.7.1",
source_dir="code",
py_version="py3",
entry_point="inference.py")
predictor = pytorch_model.deploy(initial_instance_count=1,
instance_type='local')
predictor.serializer = sagemaker.serializers.JSONSerializer()
predictor.deserializer = sagemaker.deserializers.JSONDeserializer()
result = predictor.predict("אני אוהב לעבוד באמזון")
print('result: {}'.format(result))
predictor.delete_endpoint(predictor.endpoint)
def main():
testloader = download_training_data()
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
# For local training a dummy role will be sufficient
role = 'arn:aws:iam::111111111111:role/service-role/AmazonSageMaker-ExecutionRole-20200101T000001'
print('Starting model training')
print(
'Note: if launching for the first time in local mode, container image download might take a few minutes to complete.'
)
cifar10_estimator = PyTorch(entry_point='cifar10_pytorch.py',
source_dir='./code',
role=role,
framework_version='1.7.1',
py_version='py3',
instance_count=1,
instance_type='local',
hyperparameters={
'epochs': 1,
})
cifar10_estimator.fit('file://./data/')
print('Deploying local mode endpoint')
predictor = cifar10_estimator.deploy(initial_instance_count=1,
instance_type='local')
do_inference_on_local_endpoint(predictor, testloader)
predictor.delete_endpoint(predictor.endpoint)
predictor.delete_model()
def main():
session = LocalSession()
session.config = {'local': {'local_code': True}}
role = DUMMY_IAM_ROLE
model_dir = 's3://tensorflow-script-mode-local-model-inference/model.tar.gz'
model = TensorFlowModel(
entry_point='inference.py',
source_dir='./code',
role=role,
model_data=model_dir,
framework_version='2.3.0',
)
print('Deploying endpoint in local mode')
predictor = model.deploy(
initial_instance_count=1,
instance_type='local',
)
dummy_inputs = {
'bucket_name': 'tensorflow-script-mode-local-model-inference',
'object_name': 'instances.json'
}
predictions = predictor.predict(dummy_inputs)
print("predictions: {}".format(predictions))
print('About to delete the endpoint')
predictor.delete_endpoint(predictor.endpoint_name)
predictor.delete_model()
def main():
test_loader = download_data_for_inference()
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
# For local training a dummy role will be sufficient
role = DUMMY_IAM_ROLE
model_dir = 's3://aws-ml-blog/artifacts/pytorch-script-mode-local-model-inference/model.tar.gz'
model = PyTorchModel(
role=role,
model_data=model_dir,
framework_version='1.8',
py_version='py3',
entry_point='inference.py'
)
print('Deploying endpoint in local mode')
print(
'Note: if launching for the first time in local mode, container image download might take a few minutes to complete.')
predictor = model.deploy(
initial_instance_count=1,
instance_type='local',
)
do_inference_on_local_endpoint(predictor, test_loader)
predictor.delete_endpoint(predictor.endpoint)
def main():
download_training_and_eval_data()
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
# For local training a dummy role will be sufficient
role = 'arn:aws:iam::111111111111:role/service-role/AmazonSageMaker-ExecutionRole-20200101T000001'
print('Starting model training')
mnist_estimator = TensorFlow(
entry_point='mnist_tf2.py',
role=role,
instance_count=1,
instance_type='local',
framework_version='2.3.0',
py_version='py37',
distribution={'parameter_server': {
'enabled': True
}})
mnist_estimator.fit("file://./data/")
print('Deploying local mode endpoint')
predictor = mnist_estimator.deploy(initial_instance_count=1,
instance_type='local')
do_inference_on_local_endpoint(predictor)
predictor.delete_endpoint(predictor.endpoint)
predictor.delete_model()
def get_config():
print('Will run training locally in a container image.')
session = LocalSession()
session.config = {'local': {'local_code': True}}
instance_type = 'local'
training_dataset_path = "file://./data/"
role = DUMMY_IAM_ROLE # not needed in local training
s3_data_prefix = None # not needed in local training
bucket = None # not needed in local training
config = {
's3_data_prefix': s3_data_prefix,
'sagemaker_session': session,
'bucket': bucket,
'instance_type': instance_type,
'training_dataset_path': training_dataset_path,
'role': role}
return config
def main():
session = LocalSession()
session.config = {'local': {'local_code': True}}
role = DUMMY_IAM_ROLE
model_dir = 's3://aws-ml-blog/artifacts/tensorflow-script-mode-local-model-inference/model.tar.gz'
model = TensorFlowModel(
entry_point='inference.py',
source_dir='./code',
role=role,
model_data=model_dir,
framework_version='2.3.0',
)
print('Deploying endpoint in local mode')
print(
'Note: if launching for the first time in local mode, container image download might take a few minutes to complete.'
)
predictor = model.deploy(
initial_instance_count=1,
instance_type='local',
)
print('Endpoint deployed in local mode')
dummy_inputs = {
'bucket_name':
'aws-ml-blog',
'object_name':
'artifacts/tensorflow-script-mode-local-model-inference/instances.json'
}
predictions = predictor.predict(dummy_inputs)
print("predictions: {}".format(predictions))
print('About to delete the endpoint')
predictor.delete_endpoint(predictor.endpoint_name)
def main():
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
# For local training a dummy role will be sufficient
role = DUMMY_IAM_ROLE
model_dir = 's3://aws-ml-blog/artifacts/pytorch-nlp-script-mode-local-model-inference/model.tar.gz'
test_data = pd.read_csv('./data/test_data.csv', header=None)
print(f'test_data: {test_data}')
model = PyTorchModel(role=role,
model_data=model_dir,
framework_version='1.7.1',
source_dir='code',
py_version='py3',
entry_point='inference.py')
print('Deploying endpoint in local mode')
print(
'Note: if launching for the first time in local mode, container image download might take a few minutes to complete.'
)
predictor = model.deploy(
initial_instance_count=1,
instance_type='local',
)
predictor.serializer = sagemaker.serializers.CSVSerializer()
predictor.deserializer = sagemaker.deserializers.CSVDeserializer()
predictions = predictor.predict(test_data.to_csv(header=False,
index=False))
print(f'predictions: {predictions}')
predictor.delete_endpoint(predictor.endpoint)
def get_config(mode):
assert mode is CLOUD_MODE or mode is LOCAL_MODE, f'unknown mode selected: {mode}'
if mode == CLOUD_MODE:
## REPLACE WITH A VALID IAM ROLE - START ##
role = DUMMY_IAM_ROLE
## REPLACE WITH A VALID IAM ROLE - END ##
assert role is not DUMMY_IAM_ROLE, "For cloud mode set a valid sagemaker iam role"
print('Will run training on an ML instance in AWS.')
session = sagemaker.Session()
bucket = session.default_bucket()
s3_data_prefix = 'tensorflow_script_mode_cloud_training/mnist/'
instance_type = 'ml.m5.large'
training_dataset_path = 's3://' + bucket + '/' + s3_data_prefix
else: # mode == LOCAL_MODE
print('Will run training locally in a container image.')
session = LocalSession()
session.config = {'local': {'local_code': True}}
instance_type = 'local'
training_dataset_path = "file://./data/"
role = DUMMY_IAM_ROLE # not needed in local training
s3_data_prefix = None # not needed in local training
bucket = None # not needed in local training
config = {
'mode': mode,
's3_data_prefix': s3_data_prefix,
'sagemaker_session': session,
'bucket': bucket,
'instance_type': instance_type,
'training_dataset_path': training_dataset_path,
'role': role
}
return config
from tensorflow.keras import datasets
import numpy as np
from sagemaker.tensorflow import TensorFlow
import os
from sagemaker.local import LocalSession
sagemaker_role = 'arn:aws:iam::70******AccountId:role/RoleNameHere'
sagemaker_session = LocalSession()
sagemaker_session.config = {'local': {'local_code': True}}
def sagemaker_estimator(sagemaker_role, code_entry, code_dir, hyperparameters):
sm_estimator = TensorFlow(entry_point=code_entry,
source_dir=code_dir,
role=sagemaker_role,
instance_type='local',
instance_count=1,
model_dir='/opt/ml/model',
hyperparameters=hyperparameters,
output_path='file://{}/model/'.format(
os.getcwd()),
framework_version='2.2',
py_version='py37',
script_mode=True)
return sm_estimator
def sagemaker_local_training(local_estimator, train_data_local):
local_estimator.fit({'training': train_data_local})
return local_estimator
