Merge branch 'master' into govcloud

Author: laurenyu

CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## v1.51.0 (2020-03-09)
+
+### Features
+
+ * pytorch 1.3.1 eia support
+
+### Documentation Changes
+
+ * Update Kubernetes Operator default tag
+ * improve docstring for tuner.best_estimator()
+
 ## v1.50.18.post0 (2020-03-05)
 
 ### Documentation Changes

README.rst

@@ -176,7 +176,7 @@ TensorFlow SageMaker Estimators
 
 By using TensorFlow SageMaker Estimators, you can train and host TensorFlow models on Amazon SageMaker.
 
-Supported versions of TensorFlow: ``1.4.1``, ``1.5.0``, ``1.6.0``, ``1.7.0``, ``1.8.0``, ``1.9.0``, ``1.10.0``, ``1.11.0``, ``1.12.0``, ``1.13.1``, ``1.14.0``, ``1.15.0``, ``2.0.0``.
+Supported versions of TensorFlow: ``1.4.1``, ``1.5.0``, ``1.6.0``, ``1.7.0``, ``1.8.0``, ``1.9.0``, ``1.10.0``, ``1.11.0``, ``1.12.0``, ``1.13.1``, ``1.14.0``, ``1.15.0``, ``1.15.2``, ``2.0.0``, ``2.0.1``, ``2.1.0``.
 
 Supported versions of TensorFlow for Elastic Inference: ``1.11.0``, ``1.12.0``, ``1.13.1``, ``1.14.0``.
 
@@ -208,7 +208,9 @@ PyTorch SageMaker Estimators
 
 With PyTorch SageMaker Estimators, you can train and host PyTorch models on Amazon SageMaker.
 
-Supported versions of PyTorch: ``0.4.0``, ``1.0.0``, ``1.1.0``, ``1.2.0``, ``1.3.1``.
+Supported versions of PyTorch: ``0.4.0``, ``1.0.0``, ``1.1.0``, ``1.2.0``, ``1.3.1``, ``1.4.0``.
+
+Supported versions of PyTorch for Elastic Inference: ``1.3.1``.
 
 We recommend that you use the latest supported version, because that's where we focus most of our development efforts.
 

VERSION

@@ -1 +1 @@
-1.50.19.dev0
+1.51.1.dev0

doc/amazon_sagemaker_operators_for_kubernetes.rst

@@ -349,7 +349,7 @@ that directory to your \ ``PATH``.
 
     export os="linux"
 
-    wget https://amazon-sagemaker-operator-for-k8s-us-east-1.s3.amazonaws.com/kubectl-smlogs-plugin/latest/${os}.amd64.tar.gz
+    wget https://amazon-sagemaker-operator-for-k8s-us-east-1.s3.amazonaws.com/kubectl-smlogs-plugin/v1/${os}.amd64.tar.gz
     tar xvzf ${os}.amd64.tar.gz
 
     # Move binaries to a directory in your homedir.
@@ -550,11 +550,11 @@ each region.
 +-------------+---------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------+
 | Region      | Controller Image                                                                            | Linux SMLogs                                                                                                           |
 +=============+=============================================================================================+========================================================================================================================+
-| us-east-1   | ``957583890962.dkr.ecr.us-east-1.amazonaws.com/amazon-sagemaker-operator-for-k8s:latest``   | https://amazon-sagemaker-operator-for-k8s-us-east-1.s3.amazonaws.com/kubectl-smlogs-plugin/latest/linux.amd64.tar.gz   |
+| us-east-1   | ``957583890962.dkr.ecr.us-east-1.amazonaws.com/amazon-sagemaker-operator-for-k8s:v1``       | https://amazon-sagemaker-operator-for-k8s-us-east-1.s3.amazonaws.com/kubectl-smlogs-plugin/v1/linux.amd64.tar.gz       |
 +-------------+---------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------+
-| us-east-2   | ``922499468684.dkr.ecr.us-east-2.amazonaws.com/amazon-sagemaker-operator-for-k8s:latest``   | https://amazon-sagemaker-operator-for-k8s-us-east-2.s3.amazonaws.com/kubectl-smlogs-plugin/latest/linux.amd64.tar.gz   |
+| us-east-2   | ``922499468684.dkr.ecr.us-east-2.amazonaws.com/amazon-sagemaker-operator-for-k8s:v1``       | https://amazon-sagemaker-operator-for-k8s-us-east-2.s3.amazonaws.com/kubectl-smlogs-plugin/v1/linux.amd64.tar.gz       |
 +-------------+---------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------+
-| us-west-2   | ``640106867763.dkr.ecr.us-west-2.amazonaws.com/amazon-sagemaker-operator-for-k8s:latest``   | https://amazon-sagemaker-operator-for-k8s-us-west-2.s3.amazonaws.com/kubectl-smlogs-plugin/latest/linux.amd64.tar.gz   |
+| us-west-2   | ``640106867763.dkr.ecr.us-west-2.amazonaws.com/amazon-sagemaker-operator-for-k8s:v1``       | https://amazon-sagemaker-operator-for-k8s-us-west-2.s3.amazonaws.com/kubectl-smlogs-plugin/v1/linux.amd64.tar.gz       |
 +-------------+---------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------+
-| eu-west-1   | ``613661167059.dkr.ecr.eu-west-1.amazonaws.com/amazon-sagemaker-operator-for-k8s:latest``   | https://amazon-sagemaker-operator-for-k8s-eu-west-1.s3.amazonaws.com/kubectl-smlogs-plugin/latest/linux.amd64.tar.gz   |
+| eu-west-1   | ``613661167059.dkr.ecr.eu-west-1.amazonaws.com/amazon-sagemaker-operator-for-k8s:v1``       | https://amazon-sagemaker-operator-for-k8s-eu-west-1.s3.amazonaws.com/kubectl-smlogs-plugin/v1/linux.amd64.tar.gz       |
 +-------------+---------------------------------------------------------------------------------------------+------------------------------------------------------------------------------------------------------------------------+

doc/using_pytorch.rst

@@ -6,6 +6,8 @@ With PyTorch Estimators and Models, you can train and host PyTorch models on Ama
 
 Supported versions of PyTorch: ``0.4.0``, ``1.0.0``, ``1.1.0``, ``1.2.0``, ``1.3.1``.
 
+Supported versions of PyTorch for Elastic Inference: ``1.3.1``.
+
 We recommend that you use the latest supported version, because that's where we focus most of our development efforts.
 
 You can visit the PyTorch repository at https://github.com/pytorch/pytorch.
@@ -250,6 +252,14 @@ You use the SageMaker PyTorch model server to host your PyTorch model when you c
 Estimator. The model server runs inside a SageMaker Endpoint, which your call to ``deploy`` creates.
 You can access the name of the Endpoint by the ``name`` property on the returned ``Predictor``.
 
+PyTorch on Amazon SageMaker has support for `Elastic Inference <https://docs.aws.amazon.com/sagemaker/latest/dg/ei.html>`_, which allows for inference acceleration to a hosted endpoint for a fraction of the cost of using a full GPU instance.
+In order to attach an Elastic Inference accelerator to your endpoint provide the accelerator type to ``accelerator_type`` to your ``deploy`` call.
+
+.. code:: python
+
+  predictor = pytorch_estimator.deploy(instance_type='ml.m4.xlarge',
+                                       initial_instance_count=1,
+                                       accelerator_type='ml.eia2.medium')
 
 The SageMaker PyTorch Model Server
 ==================================
@@ -291,6 +301,11 @@ It loads the model parameters from a ``model.pth`` file in the SageMaker model d
             model.load_state_dict(torch.load(f))
         return model
 
+However, if you are using PyTorch Elastic Inference, you do not have to provide a ``model_fn`` since the PyTorch serving
+container has a default one for you. But please note that if you are utilizing the default ``model_fn``, please save
+yor parameter file as ``model.pt`` instead of ``model.pth``. For more information on inference script, please refer to:
+`SageMaker PyTorch Default Inference Handler <https://github.com/aws/sagemaker-pytorch-serving-container/blob/master/src/sagemaker_pytorch_serving_container/default_inference_handler.py>`_.
+
 Serve a PyTorch Model
 ---------------------
 

src/sagemaker/algorithm.py

@@ -328,8 +328,8 @@ def transformer(
             instance_type (str): Type of EC2 instance to use, for example,
                 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in
-                a single request (default: None). Valid values: 'MULTI_RECORD'
-                and 'SINGLE_RECORD'.
+                a single request (default: None). Valid values: 'MultiRecord'
+                and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None).
                 Valid values: 'Line' or 'None'.
             output_path (str): S3 location for saving the transform result. If

src/sagemaker/estimator.py

@@ -823,8 +823,8 @@ def transformer(
             instance_type (str): Type of EC2 instance to use, for example,
                 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in
-                a single request (default: None). Valid values: 'MULTI_RECORD'
-                and 'SINGLE_RECORD'.
+                a single request (default: None). Valid values: 'MultiRecord'
+                and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None).
                 Valid values: 'Line' or 'None'.
             output_path (str): S3 location for saving the transform result. If
@@ -1895,8 +1895,8 @@ def transformer(
             instance_type (str): Type of EC2 instance to use, for example,
                 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in
-                a single request (default: None). Valid values: 'MULTI_RECORD'
-                and 'SINGLE_RECORD'.
+                a single request (default: None). Valid values: 'MultiRecord'
+                and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None).
                 Valid values: 'Line' or 'None'.
             output_path (str): S3 location for saving the transform result. If

src/sagemaker/fw_utils.py

@@ -53,7 +53,14 @@
 )
 
 VALID_PY_VERSIONS = ["py2", "py3"]
-VALID_EIA_FRAMEWORKS = ["tensorflow", "tensorflow-serving", "mxnet", "mxnet-serving"]
+VALID_EIA_FRAMEWORKS = [
+    "tensorflow",
+    "tensorflow-serving",
+    "mxnet",
+    "mxnet-serving",
+    "pytorch-serving",
+]
+PY2_RESTRICTED_EIA_FRAMEWORKS = ["pytorch-serving"]
 VALID_ACCOUNTS_BY_REGION = {"us-gov-west-1": "246785580436", "us-iso-east-1": "744548109606"}
 ASIMOV_VALID_ACCOUNTS_BY_REGION = {"us-gov-west-1": "442386744353", "us-iso-east-1": "886529160074"}
 OPT_IN_ACCOUNTS_BY_REGION = {"ap-east-1": "057415533634", "me-south-1": "724002660598"}
@@ -71,6 +78,7 @@
     "mxnet-serving-eia": "mxnet-inference-eia",
     "pytorch": "pytorch-training",
     "pytorch-serving": "pytorch-inference",
+    "pytorch-serving-eia": "pytorch-inference-eia",
 }
 
 MERGED_FRAMEWORKS_LOWEST_VERSIONS = {
@@ -82,6 +90,7 @@
     "mxnet-serving-eia": [1, 4, 1],
     "pytorch": [1, 2, 0],
     "pytorch-serving": [1, 2, 0],
+    "pytorch-serving-eia": [1, 3, 1],
 }
 
 DEBUGGER_UNSUPPORTED_REGIONS = ["us-gov-west-1", "us-iso-east-1"]
@@ -188,6 +197,7 @@ def create_image_uri(
 
     if _accelerator_type_valid_for_framework(
         framework=framework,
+        py_version=py_version,
         accelerator_type=accelerator_type,
         optimized_families=optimized_families,
     ):
@@ -240,21 +250,27 @@ def create_image_uri(
 
 
 def _accelerator_type_valid_for_framework(
-    framework, accelerator_type=None, optimized_families=None
+    framework, py_version, accelerator_type=None, optimized_families=None
 ):
     """
     Args:
         framework:
+        py_version:
         accelerator_type:
         optimized_families:
     """
     if accelerator_type is None:
         return False
 
+    if py_version == "py2" and framework in PY2_RESTRICTED_EIA_FRAMEWORKS:
+        raise ValueError(
+            "{} is not supported with Amazon Elastic Inference in Python 2.".format(framework)
+        )
+
     if framework not in VALID_EIA_FRAMEWORKS:
         raise ValueError(
             "{} is not supported with Amazon Elastic Inference. Currently only "
-            "Python-based TensorFlow and MXNet are supported.".format(framework)
+            "Python-based TensorFlow, MXNet, PyTorch are supported.".format(framework)
         )
 
     if optimized_families:

src/sagemaker/model.py

@@ -504,8 +504,8 @@ def transformer(
             instance_type (str): Type of EC2 instance to use, for example,
                 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in
-                a single request (default: None). Valid values: 'MULTI_RECORD'
-                and 'SINGLE_RECORD'.
+                a single request (default: None). Valid values: 'MultiRecord'
+                and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None).
                 Valid values: 'Line' or 'None'.
             output_path (str): S3 location for saving the transform result. If

src/sagemaker/pipeline.py

@@ -215,8 +215,8 @@ def transformer(
             instance_type (str): Type of EC2 instance to use, for example,
                 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in
-                a single request (default: None). Valid values: 'MULTI_RECORD'
-                and 'SINGLE_RECORD'.
+                a single request (default: None). Valid values: 'MultiRecord'
+                and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None).
                 Valid values: 'Line' or 'None'.
             output_path (str): S3 location for saving the transform result. If

src/sagemaker/pytorch/README.rst

@@ -6,6 +6,8 @@ With PyTorch Estimators and Models, you can train and host PyTorch models on Ama
 
 Supported versions of PyTorch: ``0.4.0``, ``1.0.0``, ``1.1.0``, ``1.2.0``, ``1.3.1``, ``1.4.0``.
 
+Supported versions of PyTorch for Elastic Inference: ``1.3.1``.
+
 We recommend that you use the latest supported version, because that's where we focus most of our development efforts.
 
 You can visit the PyTorch repository at https://github.com/pytorch/pytorch.

src/sagemaker/pytorch/model.py

@@ -136,7 +136,7 @@ def prepare_container_def(self, instance_type, accelerator_type=None):
                 For example, 'ml.p2.xlarge'.
             accelerator_type (str): The Elastic Inference accelerator type to
                 deploy to the instance for loading and making inferences to the
-                model. Currently unsupported with PyTorch.
+                model.
 
         Returns:
             dict[str, str]: A container definition object usable with the
@@ -169,7 +169,7 @@ def serving_image_uri(self, region_name, instance_type, accelerator_type=None):
                 (cpu/gpu/family-specific optimized).
             accelerator_type (str): The Elastic Inference accelerator type to
                 deploy to the instance for loading and making inferences to the
-                model. Currently unsupported with PyTorch.
+                model.
 
         Returns:
             str: The appropriate image URI based on the given parameters.

src/sagemaker/session.py

@@ -1968,7 +1968,7 @@ def transform(
             job_name (str): Name of the transform job being created.
             model_name (str): Name of the SageMaker model being used for the transform job.
             strategy (str): The strategy used to decide how to batch records in a single request.
-                Possible values are 'MULTI_RECORD' and 'SINGLE_RECORD'.
+                Possible values are 'MultiRecord' and 'SingleRecord'.
             max_concurrent_transforms (int): The maximum number of HTTP requests to be made to
                 each individual transform container at one time.
             max_payload (int): Maximum size of the payload in a single HTTP request to the

src/sagemaker/tensorflow/estimator.py

@@ -772,7 +772,7 @@ def transformer(
             instance_count (int): Number of EC2 instances to use.
             instance_type (str): Type of EC2 instance to use, for example, 'ml.c4.xlarge'.
             strategy (str): The strategy used to decide how to batch records in a single request
-                (default: None). Valid values: 'MULTI_RECORD' and 'SINGLE_RECORD'.
+                (default: None). Valid values: 'MultiRecord' and 'SingleRecord'.
             assemble_with (str): How the output is assembled (default: None). Valid values: 'Line'
                 or 'None'.
             output_path (str): S3 location for saving the transform result. If not specified,

src/sagemaker/tuner.py

@@ -760,6 +760,23 @@ def best_estimator(self, best_training_job=None):
         be deployed to an Amazon SageMaker endpoint and return a ``sagemaker.RealTimePredictor``
         object.
 
+        Args:
+            best_training_job (dict): Dictionary containing "TrainingJobName" and
+                "TrainingJobDefinitionName".
+
+                Example:
+
+                .. code:: python
+
+                    {
+                        "TrainingJobName": "my_training_job_name",
+                        "TrainingJobDefinitionName": "my_training_job_definition_name"
+                    }
+
+        Returns:
+            sagemaker.estimator.EstimatorBase: The estimator that has the best training job
+                attached.
+
         Raises:
             Exception: If there is no best training job available for the hyperparameter tuning job.
         """

src/sagemaker/workflow/airflow.py

@@ -802,8 +802,8 @@ def transform_config_from_estimator(
         model_name (str): model name (default: None). If not specified, one will
             be generated.
         strategy (str): The strategy used to decide how to batch records in a
-            single request (default: None). Valid values: 'MULTI_RECORD' and
-            'SINGLE_RECORD'.
+            single request (default: None). Valid values: 'MultiRecord' and
+            'SingleRecord'.
         assemble_with (str): How the output is assembled (default: None). Valid
             values: 'Line' or 'None'.
         output_path (str): S3 location for saving the transform result. If not

tests/data/pytorch_eia/empty_inference_script.py

@@ -0,0 +1,13 @@
+# Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License"). You
+# may not use this file except in compliance with the License. A copy of
+# the License is located at
+#
+#     http://aws.amazon.com/apache2.0/
+#
+# or in the "license" file accompanying this file. This file is
+# distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF
+# ANY KIND, either express or implied. See the License for the specific
+# language governing permissions and limitations under the License.
+# This file is intentionally left blank to invoke default model_fn and predict_fn