Merge branch 'master' into master

Author: nadiaya

CHANGELOG.md

@@ -1,5 +1,21 @@
 # Changelog
 
+## v1.62.0 (2020-06-11)
+
+### Features
+
+ * Support for multi variant endpoint invocation with target variant param
+
+### Bug Fixes and Other Changes
+
+ * Revert "feature: Support for multi variant endpoint invocation with target variant param (#1571)"
+ * make instance_type optional for prepare_container_def
+ * docs: workflows navigation
+
+### Documentation Changes
+
+ * fix typo in MXNet documentation
+
 ## v1.61.0 (2020-06-09)
 
 ### Features

VERSION

@@ -1 +1 @@
-1.61.1.dev0
+1.62.1.dev0

doc/frameworks/mxnet/using_mxnet.rst

@@ -176,7 +176,7 @@ The default serialization system generates three files:
 -  ``model-symbol.json``: The MXNet ``Module`` ``Symbol`` serialization,
    produced by invoking ``save`` on the ``symbol`` property of the
    ``Module`` being saved.
--  ``modle.params``: The MXNet ``Module`` parameters, produced by
+-  ``model.params``: The MXNet ``Module`` parameters, produced by
    invoking ``save_params`` on the ``Module`` being saved.
 
 You can provide your own save function. This is useful if you are not working with the ``Module`` API or you need special processing.

doc/index.rst

@@ -53,15 +53,14 @@ Amazon SageMaker provides implementations of some common machine learning algori
 
 
 *************
-Orchestration
+Workflows
 *************
-Orchestrate your SageMaker training and inference jobs with Kubernetes and Airflow.
+Orchestrate your SageMaker training and inference workflows with Airflow and Kubernetes.
 
 .. toctree::
-    :maxdepth: 1
+    :maxdepth: 2
 
-    kubernetes/index
-    airflow/index
+    workflows/index
 
 
 *************************

doc/airflow/index.rst renamed to doc/workflows/airflow/index.rst

@@ -0,0 +1,11 @@
+##########
+Workflows
+##########
+
+The SageMaker Python SDK supports managed training and inference for a variety of machine learning frameworks:
+
+.. toctree::
+    :maxdepth: 2
+
+    airflow/index
+    kubernetes/index

doc/airflow/sagemaker.workflow.airflow.rst renamed to doc/workflows/airflow/sagemaker.workflow.airflow.rst

@@ -34,7 +34,7 @@ def read_version():
 
 # Declare minimal set for installation
 required_packages = [
-    "boto3>=1.13.6",
+    "boto3>=1.13.24",
     "numpy>=1.9.0",
     "protobuf>=3.1",
     "scipy>=0.19.0",

doc/airflow/using_workflow.rst renamed to doc/workflows/airflow/using_workflow.rst

@@ -105,9 +105,8 @@ def __init__(
             py_version (str): Python version you want to use for executing your
                 model training code (default: 'py2'). One of 'py2' or 'py3'.
             framework_version (str): Chainer version you want to use for
-                executing your model training code. List of supported versions
-                https://github.com/aws/sagemaker-python-sdk#chainer-sagemaker-estimators.
-                If not specified, this will default to 4.1.
+                executing your model training code. If not specified, this will
+                default to 4.1.
             image_name (str): If specified, the estimator will use this image
                 for training and hosting, instead of selecting the appropriate
                 SageMaker official image based on framework_version and

doc/workflows/index.rst

@@ -15,8 +15,6 @@
 
 import logging
 
-from sagemaker import fw_utils
-
 import sagemaker
 from sagemaker.fw_utils import (
     create_image_uri,
@@ -126,7 +124,7 @@ def __init__(
         self.framework_version = framework_version or defaults.CHAINER_VERSION
         self.model_server_workers = model_server_workers
 
-    def prepare_container_def(self, instance_type, accelerator_type=None):
+    def prepare_container_def(self, instance_type=None, accelerator_type=None):
         """Return a container definition with framework configuration set in
         model environment variables.
 
@@ -143,14 +141,14 @@ def prepare_container_def(self, instance_type, accelerator_type=None):
         """
         deploy_image = self.image
         if not deploy_image:
+            if instance_type is None:
+                raise ValueError(
+                    "Must supply either an instance type (for choosing CPU vs GPU) or an image URI."
+                )
+
             region_name = self.sagemaker_session.boto_session.region_name
-            deploy_image = create_image_uri(
-                region_name,
-                self.__framework_name__,
-                instance_type,
-                self.framework_version,
-                self.py_version,
-                accelerator_type=accelerator_type,
+            deploy_image = self.serving_image_uri(
+                region_name, instance_type, accelerator_type=accelerator_type
             )
 
         deploy_key_prefix = model_code_key_prefix(self.key_prefix, self.name, deploy_image)
@@ -162,7 +160,7 @@ def prepare_container_def(self, instance_type, accelerator_type=None):
             deploy_env[MODEL_SERVER_WORKERS_PARAM_NAME.upper()] = str(self.model_server_workers)
         return sagemaker.container_def(deploy_image, self.model_data, deploy_env)
 
-    def serving_image_uri(self, region_name, instance_type):
+    def serving_image_uri(self, region_name, instance_type, accelerator_type=None):
         """Create a URI for the serving image.
 
         Args:
@@ -174,10 +172,11 @@ def serving_image_uri(self, region_name, instance_type):
             str: The appropriate image URI based on the given parameters.
 
         """
-        return fw_utils.create_image_uri(
+        return create_image_uri(
             region_name,
             self.__framework_name__,
             instance_type,
             self.framework_version,
             self.py_version,
+            accelerator_type=accelerator_type,
         )

doc/kubernetes/amazon_sagemaker_components_for_kubeflow_pipelines.rst renamed to doc/workflows/kubernetes/amazon_sagemaker_components_for_kubeflow_pipelines.rst

@@ -343,6 +343,7 @@ def invoke_endpoint(
         Accept=None,
         CustomAttributes=None,
         TargetModel=None,
+        TargetVariant=None,
     ):
         """
 
@@ -370,6 +371,9 @@ def invoke_endpoint(
         if TargetModel is not None:
             headers["X-Amzn-SageMaker-Target-Model"] = TargetModel
 
+        if TargetVariant is not None:
+            headers["X-Amzn-SageMaker-Target-Variant"] = TargetVariant
+
         r = self.http.request("POST", url, body=Body, preload_content=False, headers=headers)
 
         return {"Body": r, "ContentType": Accept}

doc/kubernetes/amazon_sagemaker_jobs.rst renamed to doc/workflows/kubernetes/amazon_sagemaker_jobs.rst

@@ -138,7 +138,7 @@ def _init_sagemaker_session_if_does_not_exist(self, instance_type):
             self.sagemaker_session = session.Session()
 
     def prepare_container_def(
-        self, instance_type, accelerator_type=None
+        self, instance_type=None, accelerator_type=None
     ):  # pylint: disable=unused-argument
         """Return a dict created by ``sagemaker.container_def()`` for deploying
         this model to a specified instance type.
@@ -166,7 +166,7 @@ def enable_network_isolation(self):
         """
         return self._enable_network_isolation
 
-    def _create_sagemaker_model(self, instance_type, accelerator_type=None, tags=None):
+    def _create_sagemaker_model(self, instance_type=None, accelerator_type=None, tags=None):
         """Create a SageMaker Model Entity
 
         Args:
@@ -807,9 +807,7 @@ def __init__(
         self.uploaded_code = None
         self.repacked_model_data = None
 
-    def prepare_container_def(
-        self, instance_type, accelerator_type=None
-    ):  # pylint disable=unused-argument
+    def prepare_container_def(self, instance_type=None, accelerator_type=None):
         """Return a container definition with framework configuration set in
         model environment variables.
 

doc/kubernetes/amazon_sagemaker_operators_for_kubernetes.rst renamed to doc/workflows/kubernetes/amazon_sagemaker_operators_for_kubernetes.rst

@@ -111,7 +111,7 @@ def __init__(
                 **kwargs
             )
 
-    def prepare_container_def(self, instance_type, accelerator_type=None):
+    def prepare_container_def(self, instance_type=None, accelerator_type=None):
         """Return a container definition set with MultiModel mode,
         model data and other parameters from the model (if available).
 

doc/kubernetes/amazon_sagemaker_operators_for_kubernetes_authentication.png renamed to doc/workflows/kubernetes/amazon_sagemaker_operators_for_kubernetes_authentication.png

@@ -126,7 +126,7 @@ def __init__(
         self.framework_version = framework_version or defaults.MXNET_VERSION
         self.model_server_workers = model_server_workers
 
-    def prepare_container_def(self, instance_type, accelerator_type=None):
+    def prepare_container_def(self, instance_type=None, accelerator_type=None):
         """Return a container definition with framework configuration set in
         model environment variables.
 
@@ -143,6 +143,11 @@ def prepare_container_def(self, instance_type, accelerator_type=None):
         """
         deploy_image = self.image
         if not deploy_image:
+            if instance_type is None:
+                raise ValueError(
+                    "Must supply either an instance type (for choosing CPU vs GPU) or an image URI."
+                )
+
             region_name = self.sagemaker_session.boto_session.region_name
             deploy_image = self.serving_image_uri(
                 region_name, instance_type, accelerator_type=accelerator_type

doc/kubernetes/index.rst renamed to doc/workflows/kubernetes/index.rst

@@ -83,7 +83,7 @@ def __init__(
         self._endpoint_config_name = self._get_endpoint_config_name()
         self._model_names = self._get_model_names()
 
-    def predict(self, data, initial_args=None, target_model=None):
+    def predict(self, data, initial_args=None, target_model=None, target_variant=None):
         """Return the inference from the specified endpoint.
 
         Args:
@@ -98,6 +98,9 @@ def predict(self, data, initial_args=None, target_model=None):
             target_model (str): S3 model artifact path to run an inference request on,
                 in case of a multi model endpoint. Does not apply to endpoints hosting
                 single model (Default: None)
+            target_variant (str): The name of the production variant to run an inference
+            request on (Default: None). Note that the ProductionVariant identifies the model
+            you want to host and the resources you want to deploy for hosting it.
 
         Returns:
             object: Inference for the given input. If a deserializer was specified when creating
@@ -106,7 +109,7 @@ def predict(self, data, initial_args=None, target_model=None):
                 as is.
         """
 
-        request_args = self._create_request_args(data, initial_args, target_model)
+        request_args = self._create_request_args(data, initial_args, target_model, target_variant)
         response = self.sagemaker_session.sagemaker_runtime_client.invoke_endpoint(**request_args)
         return self._handle_response(response)
 
@@ -123,12 +126,13 @@ def _handle_response(self, response):
         response_body.close()
         return data
 
-    def _create_request_args(self, data, initial_args=None, target_model=None):
+    def _create_request_args(self, data, initial_args=None, target_model=None, target_variant=None):
         """
         Args:
             data:
             initial_args:
             target_model:
+            target_variant:
         """
         args = dict(initial_args) if initial_args else {}
 
@@ -144,6 +148,9 @@ def _create_request_args(self, data, initial_args=None, target_model=None):
         if target_model:
             args["TargetModel"] = target_model
 
+        if target_variant:
+            args["TargetVariant"] = target_variant
+
         if self.serializer is not None:
             data = self.serializer(data)
 

doc/kubernetes/using_amazon_sagemaker_components.rst renamed to doc/workflows/kubernetes/using_amazon_sagemaker_components.rst

@@ -83,9 +83,8 @@ def __init__(
             py_version (str): Python version you want to use for executing your
                 model training code (default: 'py3'). One of 'py2' or 'py3'.
             framework_version (str): PyTorch version you want to use for
-                executing your model training code. List of supported versions
-                https://github.com/aws/sagemaker-python-sdk#pytorch-sagemaker-estimators.
-                If not specified, this will default to 0.4.
+                executing your model training code. If not specified, this will default
+                to 0.4.
             image_name (str): If specified, the estimator will use this image
                 for training and hosting, instead of selecting the appropriate
                 SageMaker official image based on framework_version and

setup.py

@@ -127,7 +127,7 @@ def __init__(
         self.framework_version = framework_version or defaults.PYTORCH_VERSION
         self.model_server_workers = model_server_workers
 
-    def prepare_container_def(self, instance_type, accelerator_type=None):
+    def prepare_container_def(self, instance_type=None, accelerator_type=None):
         """Return a container definition with framework configuration set in
         model environment variables.
 
@@ -144,6 +144,11 @@ def prepare_container_def(self, instance_type, accelerator_type=None):
         """
         deploy_image = self.image
         if not deploy_image:
+            if instance_type is None:
+                raise ValueError(
+                    "Must supply either an instance type (for choosing CPU vs GPU) or an image URI."
+                )
+
             region_name = self.sagemaker_session.boto_session.region_name
             deploy_image = self.serving_image_uri(
                 region_name, instance_type, accelerator_type=accelerator_type

src/sagemaker/chainer/estimator.py

@@ -57,8 +57,9 @@ def upload(local_path, desired_s3_uri, kms_key=None, session=None):
         """Static method that uploads a given file or directory to S3.
 
         Args:
-            local_path (str): A local path to a file or directory.
-            desired_s3_uri (str): The desired S3 uri to upload to.
+            local_path (str): Path (absolute or relative) of local file or directory to upload.
+            desired_s3_uri (str): The desired S3 location to upload to. It is the prefix to
+                which the local filename will be added.
             kms_key (str): The KMS key to use to encrypt the files.
             session (sagemaker.session.Session): Session object which
                 manages interactions with Amazon SageMaker APIs and any other

src/sagemaker/chainer/model.py

@@ -68,8 +68,7 @@ def __init__(
                 If ``source_dir`` is specified, then ``entry_point``
                 must point to a file located at the root of ``source_dir``.
             framework_version (str): Scikit-learn version you want to use for
-                executing your model training code. List of supported versions
-                https://github.com/aws/sagemaker-python-sdk#sklearn-sagemaker-estimators
+                executing your model training code.
             source_dir (str): Path (absolute, relative or an S3 URI) to a directory
                 with any other training source code dependencies aside from the entry
                 point file (default: None). If ``source_dir`` is an S3 URI, it must