Merge branch 'zwei' into require-framework-version-tensorflow

Author: metrizable

doc/frameworks/pytorch/using_pytorch.rst

@@ -154,7 +154,8 @@ directories ('train' and 'test').
     pytorch_estimator = PyTorch('pytorch-train.py',
                                 train_instance_type='ml.p3.2xlarge',
                                 train_instance_count=1,
-                                framework_version='1.0.0',
+                                framework_version='1.5.0',
+                                py_version='py3',
                                 hyperparameters = {'epochs': 20, 'batch-size': 64, 'learning-rate': 0.1})
     pytorch_estimator.fit({'train': 's3://my-data-bucket/path/to/my/training/data',
                            'test': 's3://my-data-bucket/path/to/my/test/data'})
@@ -247,7 +248,8 @@ operation.
     pytorch_estimator = PyTorch(entry_point='train_and_deploy.py',
                                 train_instance_type='ml.p3.2xlarge',
                                 train_instance_count=1,
-                                framework_version='1.0.0')
+                                framework_version='1.5.0',
+                                py_version='py3')
     pytorch_estimator.fit('s3://my_bucket/my_training_data/')
 
     # Deploy my estimator to a SageMaker Endpoint and get a Predictor

src/sagemaker/amazon/factorization_machines.py

@@ -310,8 +310,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
         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,
+            model_data,
             role,
             predictor_cls=FactorizationMachinesPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/ipinsights.py

@@ -216,8 +216,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
         )
 
         super(IPInsightsModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=IPInsightsPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/kmeans.py

@@ -241,8 +241,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
         repo = "{}:{}".format(KMeans.repo_name, KMeans.repo_version)
         image = "{}/{}".format(registry(sagemaker_session.boto_session.region_name), repo)
         super(KMeansModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=KMeansPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/knn.py

@@ -231,8 +231,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
             registry(sagemaker_session.boto_session.region_name, KNN.repo_name), repo
         )
         super(KNNModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=KNNPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/lda.py

@@ -215,8 +215,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
             registry(sagemaker_session.boto_session.region_name, LDA.repo_name), repo
         )
         super(LDAModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=LDAPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/linear_learner.py

@@ -474,8 +474,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
         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,
+            model_data,
             role,
             predictor_cls=LinearLearnerPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/ntm.py

@@ -245,8 +245,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
             registry(sagemaker_session.boto_session.region_name, NTM.repo_name), repo
         )
         super(NTMModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=NTMPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/object2vec.py

@@ -355,8 +355,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
             registry(sagemaker_session.boto_session.region_name, Object2Vec.repo_name), repo
         )
         super(Object2VecModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=RealTimePredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/pca.py

@@ -225,8 +225,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
         repo = "{}:{}".format(PCA.repo_name, PCA.repo_version)
         image = "{}/{}".format(registry(sagemaker_session.boto_session.region_name), repo)
         super(PCAModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=PCAPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/amazon/randomcutforest.py

@@ -206,8 +206,8 @@ def __init__(self, model_data, role, sagemaker_session=None, **kwargs):
             registry(sagemaker_session.boto_session.region_name, RandomCutForest.repo_name), repo
         )
         super(RandomCutForestModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=RandomCutForestPredictor,
             sagemaker_session=sagemaker_session,

src/sagemaker/estimator.py

@@ -1398,8 +1398,8 @@ def predict_wrapper(endpoint, session):
             kwargs["enable_network_isolation"] = self.enable_network_isolation()
 
         return Model(
-            self.model_data,
             image or self.train_image(),
+            self.model_data,
             role,
             vpc_config=self.get_vpc_config(vpc_config_override),
             sagemaker_session=self.sagemaker_session,

src/sagemaker/model.py

@@ -60,8 +60,8 @@ class Model(object):
 
     def __init__(
         self,
-        model_data,
         image,
+        model_data=None,
         role=None,
         predictor_cls=None,
         env=None,
@@ -74,9 +74,9 @@ def __init__(
         """Initialize an SageMaker ``Model``.
 
         Args:
-            model_data (str): The S3 location of a SageMaker model data
-                ``.tar.gz`` file.
             image (str): A Docker image URI.
+            model_data (str): The S3 location of a SageMaker model data
+                ``.tar.gz`` file (default: None).
             role (str): An AWS IAM role (either name or full ARN). The Amazon
                 SageMaker training jobs and APIs that create Amazon SageMaker
                 endpoints use this role to access training data and model
@@ -361,6 +361,8 @@ def compile(
             )
         if job_name is None:
             raise ValueError("You must provide a compilation job name")
+        if self.model_data is None:
+            raise ValueError("You must provide an S3 path to the compressed model artifacts.")
 
         framework = framework.upper()
         framework_version = self._get_framework_version() or framework_version
@@ -778,8 +780,8 @@ def __init__(
             :class:`~sagemaker.model.Model`.
         """
         super(FrameworkModel, self).__init__(
-            model_data,
             image,
+            model_data,
             role,
             predictor_cls=predictor_cls,
             env=env,

src/sagemaker/multidatamodel.py

@@ -103,8 +103,8 @@ def __init__(
         # Set the ``Model`` parameters if the model parameter is not specified
         if not self.model:
             super(MultiDataModel, self).__init__(
-                self.model_data_prefix,
                 image,
+                self.model_data_prefix,
                 role,
                 name=self.name,
                 sagemaker_session=self.sagemaker_session,

src/sagemaker/pytorch/estimator.py

@@ -19,9 +19,9 @@
 from sagemaker.fw_utils import (
     framework_name_from_image,
     framework_version_from_tag,
-    empty_framework_version_warning,
-    python_deprecation_warning,
     is_version_equal_or_higher,
+    python_deprecation_warning,
+    validate_version_or_image_args,
 )
 from sagemaker.pytorch import defaults
 from sagemaker.pytorch.model import PyTorchModel
@@ -40,10 +40,10 @@ class PyTorch(Framework):
     def __init__(
         self,
         entry_point,
+        framework_version=None,
+        py_version=None,
         source_dir=None,
         hyperparameters=None,
-        py_version=defaults.PYTHON_VERSION,
-        framework_version=None,
         image_name=None,
         **kwargs
     ):
@@ -69,6 +69,13 @@ def __init__(
                 file which should be executed as the entry point to training.
                 If ``source_dir`` is specified, then ``entry_point``
                 must point to a file located at the root of ``source_dir``.
+            framework_version (str): PyTorch version you want to use for
+                executing your model training code. Defaults to ``None``. Required unless
+                ``image_name`` is provided. List of supported versions:
+                https://github.com/aws/sagemaker-python-sdk#pytorch-sagemaker-estimators.
+            py_version (str): Python version you want to use for executing your
+                model training code. One of 'py2' or 'py3'. Defaults to ``None``. Required
+                unless ``image_name`` is provided.
             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
@@ -80,12 +87,6 @@ def __init__(
                 SageMaker. For convenience, this accepts other types for keys
                 and values, but ``str()`` will be called to convert them before
                 training.
-            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.
             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
@@ -95,6 +96,9 @@ def __init__(
                     * ``123412341234.dkr.ecr.us-west-2.amazonaws.com/my-custom-image:1.0``
                     * ``custom-image:latest``
 
+                If ``framework_version`` or ``py_version`` are ``None``, then
+                ``image_name`` is required. If also ``None``, then a ``ValueError``
+                will be raised.
             **kwargs: Additional kwargs passed to the :class:`~sagemaker.estimator.Framework`
                 constructor.
 
@@ -104,28 +108,25 @@ def __init__(
             :class:`~sagemaker.estimator.Framework` and
             :class:`~sagemaker.estimator.EstimatorBase`.
         """
-        if framework_version is None:
+        validate_version_or_image_args(framework_version, py_version, image_name)
+        if py_version == "py2":
             logger.warning(
-                empty_framework_version_warning(defaults.PYTORCH_VERSION, self.LATEST_VERSION)
+                python_deprecation_warning(self.__framework_name__, defaults.LATEST_PY2_VERSION)
             )
-        self.framework_version = framework_version or defaults.PYTORCH_VERSION
+        self.framework_version = framework_version
+        self.py_version = py_version
 
         if "enable_sagemaker_metrics" not in kwargs:
             # enable sagemaker metrics for PT v1.3 or greater:
-            if is_version_equal_or_higher([1, 3], self.framework_version):
+            if self.framework_version and is_version_equal_or_higher(
+                [1, 3], self.framework_version
+            ):
                 kwargs["enable_sagemaker_metrics"] = True
 
         super(PyTorch, self).__init__(
             entry_point, source_dir, hyperparameters, image_name=image_name, **kwargs
         )
 
-        if py_version == "py2":
-            logger.warning(
-                python_deprecation_warning(self.__framework_name__, defaults.LATEST_PY2_VERSION)
-            )
-
-        self.py_version = py_version
-
     def create_model(
         self,
         model_server_workers=None,
@@ -177,12 +178,12 @@ def create_model(
             self.model_data,
             role or self.role,
             entry_point or self.entry_point,
+            framework_version=self.framework_version,
+            py_version=self.py_version,
             source_dir=(source_dir or self._model_source_dir()),
             enable_cloudwatch_metrics=self.enable_cloudwatch_metrics,
             container_log_level=self.container_log_level,
             code_location=self.code_location,
-            py_version=self.py_version,
-            framework_version=self.framework_version,
             model_server_workers=model_server_workers,
             sagemaker_session=self.sagemaker_session,
             vpc_config=self.get_vpc_config(vpc_config_override),
@@ -210,15 +211,19 @@ class constructor
         image_name = init_params.pop("image")
         framework, py_version, tag, _ = framework_name_from_image(image_name)
 
+        if tag is None:
+            framework_version = None
+        else:
+            framework_version = framework_version_from_tag(tag)
+        init_params["framework_version"] = framework_version
+        init_params["py_version"] = py_version
+
         if not framework:
             # If we were unable to parse the framework name from the image it is not one of our
             # officially supported images, in this case just add the image to the init params.
             init_params["image_name"] = image_name
             return init_params
 
-        init_params["py_version"] = py_version
-        init_params["framework_version"] = framework_version_from_tag(tag)
-
         training_job_name = init_params["base_job_name"]
 
         if framework != cls.__framework_name__:

src/sagemaker/pytorch/model.py

@@ -21,7 +21,7 @@
     create_image_uri,
     model_code_key_prefix,
     python_deprecation_warning,
-    empty_framework_version_warning,
+    validate_version_or_image_args,
 )
 from sagemaker.model import FrameworkModel, MODEL_SERVER_WORKERS_PARAM_NAME
 from sagemaker.pytorch import defaults
@@ -66,9 +66,9 @@ def __init__(
         model_data,
         role,
         entry_point,
-        image=None,
-        py_version=defaults.PYTHON_VERSION,
         framework_version=None,
+        py_version=None,
+        image=None,
         predictor_cls=PyTorchPredictor,
         model_server_workers=None,
         **kwargs
@@ -87,12 +87,16 @@ def __init__(
                 file which should be executed as the entry point to model
                 hosting. If ``source_dir`` is specified, then ``entry_point``
                 must point to a file located at the root of ``source_dir``.
-            image (str): A Docker image URI (default: None). If not specified, a
-                default image for PyTorch will be used.
-            py_version (str): Python version you want to use for executing your
-                model training code (default: 'py3').
             framework_version (str): PyTorch version you want to use for
-                executing your model training code.
+                executing your model training code. Defaults to None. Required
+                unless ``image`` is provided.
+            py_version (str): Python version you want to use for executing your
+                model training code. Defaults to ``None``. Required unless
+                ``image`` is provided.
+            image (str): A Docker image URI (default: None). If not specified, a
+                default image for PyTorch will be used. If ``framework_version``
+                or ``py_version`` are ``None``, then ``image`` is required. If
+                also ``None``, then a ``ValueError`` will be raised.
             predictor_cls (callable[str, sagemaker.session.Session]): A function
                 to call to create a predictor with an endpoint name and
                 SageMaker ``Session``. If specified, ``deploy()`` returns the
@@ -109,22 +113,18 @@ def __init__(
             :class:`~sagemaker.model.FrameworkModel` and
             :class:`~sagemaker.model.Model`.
         """
-        super(PyTorchModel, self).__init__(
-            model_data, image, role, entry_point, predictor_cls=predictor_cls, **kwargs
-        )
-
-        if py_version == "py2":
+        validate_version_or_image_args(framework_version, py_version, image)
+        if py_version and py_version == "py2":
             logger.warning(
                 python_deprecation_warning(self.__framework_name__, defaults.LATEST_PY2_VERSION)
             )
+        self.framework_version = framework_version
+        self.py_version = py_version
 
-        if framework_version is None:
-            logger.warning(
-                empty_framework_version_warning(defaults.PYTORCH_VERSION, defaults.LATEST_VERSION)
-            )
+        super(PyTorchModel, self).__init__(
+            model_data, image, role, entry_point, predictor_cls=predictor_cls, **kwargs
+        )
 
-        self.py_version = py_version
-        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):