Merge branch 'master' of https://github.com/aws/sagemaker-python-sdk

Author: aaronmarkham

CHANGELOG.md

@@ -1,5 +1,43 @@
 # Changelog
 
+## v1.55.3 (2020-04-08)
+
+### Bug Fixes and Other Changes
+
+ * remove .strip() from batch transform
+ * allow model with network isolation when creating a Transformer from an Estimator
+ * add enable_network_isolation to EstimatorBase
+
+## v1.55.2 (2020-04-07)
+
+### Bug Fixes and Other Changes
+
+ * use .format instead of os.path.join for Processing S3 paths.
+
+### Testing and Release Infrastructure
+
+ * use m5.xlarge instances for "ap-northeast-1" region integ tests.
+
+## v1.55.1 (2020-04-06)
+
+### Bug Fixes and Other Changes
+
+ * correct local mode behavior for CN regions
+
+## v1.55.0.post0 (2020-04-06)
+
+### Documentation Changes
+
+ * fix documentation to provide working example.
+ * add documentation for XGBoost
+ * Correct comment in SKLearn Estimator about default Python version
+ * document inferentia supported version
+ * Merge Amazon Sagemaker Operators for Kubernetes and Kubernetes Jobs pages
+
+### Testing and Release Infrastructure
+
+ * turn on warnings as errors for docs builds
+
 ## v1.55.0 (2020-03-31)
 
 ### Features

VERSION

@@ -1 +1 @@
-1.55.1.dev0
+1.55.4.dev0

doc/using_tf.rst

@@ -267,19 +267,19 @@ Training with ``MPI`` is configured by specifying following fields in ``distribu
   command executed by SageMaker to launch distributed horovod training.
 
 
-In the below example we create an estimator to launch Horovod distributed training with 2 processes on one host:
+In the below example we create an estimator to launch Horovod distributed training with 4 processes on one host:
 
 .. code:: python
 
     from sagemaker.tensorflow import TensorFlow
 
     tf_estimator = TensorFlow(entry_point='tf-train.py', role='SageMakerRole',
-                              train_instance_count=1, train_instance_type='ml.p2.xlarge',
-                              framework_version='1.12', py_version='py3',
+                              train_instance_count=1, train_instance_type='ml.p3.8xlarge',
+                              framework_version='2.1.0', py_version='py3',
                               distributions={
                                   'mpi': {
                                       'enabled': True,
-                                      'processes_per_host': 2,
+                                      'processes_per_host': 4,
                                       'custom_mpi_options': '--NCCL_DEBUG INFO'
                                   }
                               })

src/sagemaker/amazon/amazon_estimator.py

@@ -82,14 +82,17 @@ def __init__(
             :class:`~sagemaker.estimator.EstimatorBase`.
         """
         super(AmazonAlgorithmEstimatorBase, self).__init__(
-            role, train_instance_count, train_instance_type, **kwargs
+            role,
+            train_instance_count,
+            train_instance_type,
+            enable_network_isolation=enable_network_isolation,
+            **kwargs
         )
 
         data_location = data_location or "s3://{}/sagemaker-record-sets/".format(
             self.sagemaker_session.default_bucket()
         )
         self._data_location = data_location
-        self._enable_network_isolation = enable_network_isolation
 
     def train_image(self):
         """Placeholder docstring"""
@@ -101,14 +104,6 @@ def hyperparameters(self):
         """Placeholder docstring"""
         return hp.serialize_all(self)
 
-    def enable_network_isolation(self):
-        """If this Estimator can use network isolation when running.
-
-        Returns:
-            bool: Whether this Estimator can use network isolation or not.
-        """
-        return self._enable_network_isolation
-
     @property
     def data_location(self):
         """Placeholder docstring"""

src/sagemaker/chainer/estimator.py

@@ -206,13 +206,15 @@ def create_model(
         if "image" not in kwargs:
             kwargs["image"] = self.image_name
 
+        if "name" not in kwargs:
+            kwargs["name"] = self._current_job_name
+
         return ChainerModel(
             self.model_data,
             role or self.role,
             entry_point or self.entry_point,
             source_dir=(source_dir or self._model_source_dir()),
             enable_cloudwatch_metrics=self.enable_cloudwatch_metrics,
-            name=self._current_job_name,
             container_log_level=self.container_log_level,
             code_location=self.code_location,
             py_version=self.py_version,

src/sagemaker/estimator.py

@@ -98,6 +98,7 @@ def __init__(
         debugger_hook_config=None,
         tensorboard_output_config=None,
         enable_sagemaker_metrics=None,
+        enable_network_isolation=False,
     ):
         """Initialize an ``EstimatorBase`` instance.
 
@@ -199,6 +200,11 @@ def __init__(
                 Series. For more information see:
                 https://docs.aws.amazon.com/sagemaker/latest/dg/API_AlgorithmSpecification.html#SageMaker-Type-AlgorithmSpecification-EnableSageMakerMetricsTimeSeries
                 (default: ``None``).
+            enable_network_isolation (bool): Specifies whether container will
+                run in network isolation mode (default: ``False``). Network
+                isolation mode restricts the container access to outside networks
+                (such as the Internet). The container does not make any inbound or
+                outbound network calls. Also known as Internet-free mode.
         """
         self.role = role
         self.train_instance_count = train_instance_count
@@ -260,6 +266,7 @@ def __init__(
         self.collection_configs = None
 
         self.enable_sagemaker_metrics = enable_sagemaker_metrics
+        self._enable_network_isolation = enable_network_isolation
 
     @abstractmethod
     def train_image(self):
@@ -290,7 +297,7 @@ def enable_network_isolation(self):
         Returns:
             bool: Whether this Estimator needs network isolation or not.
         """
-        return False
+        return self._enable_network_isolation
 
     def prepare_workflow_for_training(self, job_name=None):
         """Calls _prepare_for_training. Used when setting up a workflow.
@@ -818,6 +825,8 @@ def transformer(
         role=None,
         volume_kms_key=None,
         vpc_config_override=vpc_utils.VPC_CONFIG_DEFAULT,
+        enable_network_isolation=None,
+        model_name=None,
     ):
         """Return a ``Transformer`` that uses a SageMaker Model based on the
         training job. It reuses the SageMaker Session and base job name used by
@@ -856,8 +865,20 @@ def transformer(
             vpc_config_override (dict[str, list[str]]): Optional override for the
                 VpcConfig set on the model.
                 Default: use subnets and security groups from this Estimator.
+
                 * 'Subnets' (list[str]): List of subnet ids.
                 * 'SecurityGroupIds' (list[str]): List of security group ids.
+
+            enable_network_isolation (bool): Specifies whether container will
+                run in network isolation mode. Network isolation mode restricts
+                the container access to outside networks (such as the internet).
+                The container does not make any inbound or outbound network
+                calls. If True, a channel named "code" will be created for any
+                user entry script for inference. Also known as Internet-free mode.
+                If not specified, this setting is taken from the estimator's
+                current configuration.
+            model_name (str): Name to use for creating an Amazon SageMaker
+                model. If not specified, the name of the training job is used.
         """
         tags = tags or self.tags
 
@@ -866,11 +887,16 @@ def transformer(
                 "No finished training job found associated with this estimator. Please make sure "
                 "this estimator is only used for building workflow config"
             )
-            model_name = self._current_job_name
+            model_name = model_name or self._current_job_name
         else:
-            model_name = self.latest_training_job.name
+            model_name = model_name or self.latest_training_job.name
+            if enable_network_isolation is None:
+                enable_network_isolation = self.enable_network_isolation()
+
             model = self.create_model(
-                vpc_config_override=vpc_config_override, model_kms_key=self.output_kms_key
+                vpc_config_override=vpc_config_override,
+                model_kms_key=self.output_kms_key,
+                enable_network_isolation=enable_network_isolation,
             )
 
             # not all create_model() implementations have the same kwargs
@@ -1219,21 +1245,17 @@ def __init__(
                 checkpoints will be provided under `/opt/ml/checkpoints/`.
                 (default: ``None``).
             enable_network_isolation (bool): Specifies whether container will
-                run in network isolation mode. Network isolation mode restricts
-                the container access to outside networks (such as the Internet).
-                The container does not make any inbound or outbound network
-                calls. If ``True``, a channel named "code" will be created for any
-                user entry script for training. The user entry script, files in
-                source_dir (if specified), and dependencies will be uploaded in
-                a tar to S3. Also known as internet-free mode (default: ``False``).
+                run in network isolation mode (default: ``False``). Network
+                isolation mode restricts the container access to outside networks
+                (such as the Internet). The container does not make any inbound or
+                outbound network calls. Also known as Internet-free mode.
             enable_sagemaker_metrics (bool): enable SageMaker Metrics Time
                 Series. For more information see:
                 https://docs.aws.amazon.com/sagemaker/latest/dg/API_AlgorithmSpecification.html#SageMaker-Type-AlgorithmSpecification-EnableSageMakerMetricsTimeSeries
                 (default: ``None``).
         """
         self.image_name = image_name
         self.hyperparam_dict = hyperparameters.copy() if hyperparameters else {}
-        self._enable_network_isolation = enable_network_isolation
         super(Estimator, self).__init__(
             role,
             train_instance_count,
@@ -1261,16 +1283,9 @@ def __init__(
             debugger_hook_config=debugger_hook_config,
             tensorboard_output_config=tensorboard_output_config,
             enable_sagemaker_metrics=enable_sagemaker_metrics,
+            enable_network_isolation=enable_network_isolation,
         )
 
-    def enable_network_isolation(self):
-        """If this Estimator can use network isolation when running.
-
-        Returns:
-            bool: Whether this Estimator can use network isolation or not.
-        """
-        return self._enable_network_isolation
-
     def train_image(self):
         """Returns the docker image to use for training.
 
@@ -1358,14 +1373,16 @@ def predict_wrapper(endpoint, session):
 
         role = role or self.role
 
+        if "enable_network_isolation" not in kwargs:
+            kwargs["enable_network_isolation"] = self.enable_network_isolation()
+
         return Model(
             self.model_data,
             image or self.train_image(),
             role,
             vpc_config=self.get_vpc_config(vpc_config_override),
             sagemaker_session=self.sagemaker_session,
             predictor_cls=predictor_cls,
-            enable_network_isolation=self.enable_network_isolation(),
             **kwargs
         )
 
@@ -1498,15 +1515,15 @@ def __init__(
                     >>>     |------ train.py
                     >>>     |------ common
                     >>>     |------ virtual-env
+
             enable_network_isolation (bool): Specifies whether container will
                 run in network isolation mode. Network isolation mode restricts
                 the container access to outside networks (such as the internet).
                 The container does not make any inbound or outbound network
                 calls. If True, a channel named "code" will be created for any
                 user entry script for training. The user entry script, files in
                 source_dir (if specified), and dependencies will be uploaded in
-                a tar to S3. Also known as internet-free mode (default: `False`
-                ).
+                a tar to S3. Also known as internet-free mode (default: `False`).
             git_config (dict[str, str]): Git configurations used for cloning
                 files, including ``repo``, ``branch``, ``commit``,
                 ``2FA_enabled``, ``username``, ``password`` and ``token``. The
@@ -1579,7 +1596,7 @@ def __init__(
             You can find additional parameters for initializing this class at
             :class:`~sagemaker.estimator.EstimatorBase`.
         """
-        super(Framework, self).__init__(**kwargs)
+        super(Framework, self).__init__(enable_network_isolation=enable_network_isolation, **kwargs)
         if entry_point.startswith("s3://"):
             raise ValueError(
                 "Invalid entry point script: {}. Must be a path to a local file.".format(
@@ -1599,7 +1616,6 @@ def __init__(
         self.container_log_level = container_log_level
         self.code_location = code_location
         self.image_name = image_name
-        self._enable_network_isolation = enable_network_isolation
 
         self.uploaded_code = None
 
@@ -1608,14 +1624,6 @@ def __init__(
         self.checkpoint_local_path = checkpoint_local_path
         self.enable_sagemaker_metrics = enable_sagemaker_metrics
 
-    def enable_network_isolation(self):
-        """Return True if this Estimator can use network isolation to run.
-
-        Returns:
-            bool: Whether this Estimator can use network isolation or not.
-        """
-        return self._enable_network_isolation
-
     def _prepare_for_training(self, job_name=None):
         """Set hyperparameters needed for training. This method will also
         validate ``source_dir``.
@@ -1891,6 +1899,8 @@ def transformer(
         volume_kms_key=None,
         entry_point=None,
         vpc_config_override=vpc_utils.VPC_CONFIG_DEFAULT,
+        enable_network_isolation=None,
+        model_name=None,
     ):
         """Return a ``Transformer`` that uses a SageMaker Model based on the
         training job. It reuses the SageMaker Session and base job name used by
@@ -1935,9 +1945,21 @@ def transformer(
             vpc_config_override (dict[str, list[str]]): Optional override for
                 the VpcConfig set on the model.
                 Default: use subnets and security groups from this Estimator.
+
                 * 'Subnets' (list[str]): List of subnet ids.
                 * 'SecurityGroupIds' (list[str]): List of security group ids.
 
+            enable_network_isolation (bool): Specifies whether container will
+                run in network isolation mode. Network isolation mode restricts
+                the container access to outside networks (such as the internet).
+                The container does not make any inbound or outbound network
+                calls. If True, a channel named "code" will be created for any
+                user entry script for inference. Also known as Internet-free mode.
+                If not specified, this setting is taken from the estimator's
+                current configuration.
+            model_name (str): Name to use for creating an Amazon SageMaker
+                model. If not specified, the name of the training job is used.
+
         Returns:
             sagemaker.transformer.Transformer: a ``Transformer`` object that can be used to start a
                 SageMaker Batch Transform job.
@@ -1946,12 +1968,17 @@ def transformer(
         tags = tags or self.tags
 
         if self.latest_training_job is not None:
+            if enable_network_isolation is None:
+                enable_network_isolation = self.enable_network_isolation()
+
             model = self.create_model(
                 role=role,
                 model_server_workers=model_server_workers,
                 entry_point=entry_point,
                 vpc_config_override=vpc_config_override,
                 model_kms_key=self.output_kms_key,
+                enable_network_isolation=enable_network_isolation,
+                name=model_name,
             )
             model._create_sagemaker_model(instance_type, tags=tags)
 
@@ -1964,7 +1991,7 @@ def transformer(
                 "No finished training job found associated with this estimator. Please make sure "
                 "this estimator is only used for building workflow config"
             )
-            model_name = self._current_job_name
+            model_name = model_name or self._current_job_name
             transform_env = env or {}
 
         return Transformer(

src/sagemaker/local/entities.py

@@ -356,7 +356,7 @@ def _perform_batch_inference(self, input_data, output_data, **kwargs):
                     )
 
                     response_body = response["Body"]
-                    data = response_body.read().strip()
+                    data = response_body.read()
                     response_body.close()
                     f.write(data)
                     if "AssembleWith" in output_data and output_data["AssembleWith"] == "Line":