Estimators

A high level interface for SageMaker training

class sagemaker.estimator.EstimatorBase(role, train_instance_count, train_instance_type, train_volume_size=30, train_volume_kms_key=None, train_max_run=86400, input_mode='File', output_path=None, output_kms_key=None, base_job_name=None, sagemaker_session=None, tags=None, subnets=None, security_group_ids=None, model_uri=None, model_channel_name='model', metric_definitions=None, encrypt_inter_container_traffic=False)

Bases: object

Handle end-to-end Amazon SageMaker training and deployment tasks.

For introduction to model training and deployment, see http://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-training.html

Subclasses must define a way to determine what image to use for training, what hyperparameters to use, and how to create an appropriate predictor instance.

Initialize an EstimatorBase instance.

Parameters:
  • 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 artifacts. After the endpoint is created, the inference code might use the IAM role, if it needs to access an AWS resource.
  • train_instance_count (int) – Number of Amazon EC2 instances to use for training.
  • train_instance_type (str) – Type of EC2 instance to use for training, for example, ‘ml.c4.xlarge’.
  • train_volume_size (int) – Size in GB of the EBS volume to use for storing input data during training (default: 30). Must be large enough to store training data if File Mode is used (which is the default).
  • train_volume_kms_key (str) – Optional. KMS key ID for encrypting EBS volume attached to the training instance (default: None).
  • train_max_run (int) – Timeout in seconds for training (default: 24 * 60 * 60). After this amount of time Amazon SageMaker terminates the job regardless of its current status.
  • input_mode (str) – The input mode that the algorithm supports (default: ‘File’). Valid modes: ‘File’ - Amazon SageMaker copies the training dataset from the S3 location to a local directory. ‘Pipe’ - Amazon SageMaker streams data directly from S3 to the container via a Unix-named pipe. This argument can be overriden on a per-channel basis using sagemaker.session.s3_input.input_mode.
  • output_path (str) – S3 location for saving the trainig result (model artifacts and output files). If not specified, results are stored to a default bucket. If the bucket with the specific name does not exist, the estimator creates the bucket during the fit() method execution.
  • output_kms_key (str) – Optional. KMS key ID for encrypting the training output (default: None).
  • base_job_name (str) – Prefix for training job name when the fit() method launches. If not specified, the estimator generates a default job name, based on the training image name and current timestamp.
  • sagemaker_session (sagemaker.session.Session) – Session object which manages interactions with Amazon SageMaker APIs and any other AWS services needed. If not specified, the estimator creates one using the default AWS configuration chain.
  • tags (list[dict]) – List of tags for labeling a training job. For more, see https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
  • subnets (list[str]) – List of subnet ids. If not specified training job will be created without VPC config.
  • security_group_ids (list[str]) – List of security group ids. If not specified training job will be created without VPC config.
  • model_uri (str) –

    URI where a pre-trained model is stored, either locally or in S3 (default: None). If specified, the estimator will create a channel pointing to the model so the training job can download it. This model can be a ‘model.tar.gz’ from a previous training job, or other artifacts coming from a different source.

    In local mode, this should point to the path in which the model is located and not the file itself, as local Docker containers will try to mount the URI as a volume.

    More information: https://docs.aws.amazon.com/sagemaker/latest/dg/cdf-training.html#td-deserialization

  • model_channel_name (str) – Name of the channel where ‘model_uri’ will be downloaded (default: ‘model’).
  • metric_definitions (list[dict]) – A list of dictionaries that defines the metric(s) used to evaluate the training jobs. Each dictionary contains two keys: ‘Name’ for the name of the metric, and ‘Regex’ for the regular expression used to extract the metric from the logs. This should be defined only for jobs that don’t use an Amazon algorithm.
  • encrypt_inter_container_traffic (bool) – Specifies whether traffic between training containers is encrypted for the training job (default: False).
train_image()

Return the Docker image to use for training.

The fit() method, which does the model training, calls this method to find the image to use for model training.

Returns:The URI of the Docker image.
Return type:str
hyperparameters()

Return the hyperparameters as a dictionary to use for training.

The fit() method, which trains the model, calls this method to find the hyperparameters.

Returns:The hyperparameters.
Return type:dict[str, str]
enable_network_isolation()

Return True if this Estimator will need network isolation to run.

Returns:Whether this Estimator needs network isolation or not.
Return type:bool
fit(inputs=None, wait=True, logs=True, job_name=None)

Train a model using the input training dataset.

The API calls the Amazon SageMaker CreateTrainingJob API to start model training. The API uses configuration you provided to create the estimator and the specified input training data to send the CreatingTrainingJob request to Amazon SageMaker.

This is a synchronous operation. After the model training successfully completes, you can call the deploy() method to host the model using the Amazon SageMaker hosting services.

Parameters:
  • inputs (str or dict or sagemaker.session.s3_input) –

    Information about the training data. This can be one of three types:

    • (str) the S3 location where training data is saved.
    • (dict[str, str] or dict[str, sagemaker.session.s3_input]) If using multiple channels for
      training data, you can specify a dict mapping channel names to strings or s3_input() objects.
    • (sagemaker.session.s3_input) - channel configuration for S3 data sources that can provide
      additional information as well as the path to the training dataset. See sagemaker.session.s3_input() for full details.
  • wait (bool) – Whether the call should wait until the job completes (default: True).
  • logs (bool) – Whether to show the logs produced by the job. Only meaningful when wait is True (default: True).
  • job_name (str) – Training job name. If not specified, the estimator generates a default job name, based on the training image name and current timestamp.
compile_model(target_instance_family, input_shape, output_path, framework=None, framework_version=None, compile_max_run=300, tags=None, **kwargs)

Compile a Neo model using the input model.

Parameters:
  • target_instance_family (str) – Identifies the device that you want to run your model after compilation, for example: ml_c5. Allowed strings are: ml_c5, ml_m5, ml_c4, ml_m4, jetsontx1, jetsontx2, ml_p2, ml_p3, deeplens, rasp3b
  • input_shape (dict) – Specifies the name and shape of the expected inputs for your trained model in json dictionary form, for example: {‘data’:[1,3,1024,1024]}, or {‘var1’: [1,1,28,28], ‘var2’:[1,1,28,28]}
  • output_path (str) – Specifies where to store the compiled model
  • framework (str) – The framework that is used to train the original model. Allowed values: ‘mxnet’, ‘tensorflow’, ‘pytorch’, ‘onnx’, ‘xgboost’
  • framework_version (str) – The version of the framework
  • compile_max_run (int) – Timeout in seconds for compilation (default: 3 * 60). After this amount of time Amazon SageMaker Neo terminates the compilation job regardless of its current status.
  • tags (list[dict]) – List of tags for labeling a compilation job. For more, see https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A SageMaker Model object. See Model() for full details.

Return type:

sagemaker.model.Model

classmethod attach(training_job_name, sagemaker_session=None, model_channel_name='model')

Attach to an existing training job.

Create an Estimator bound to an existing training job, each subclass is responsible to implement _prepare_init_params_from_job_description() as this method delegates the actual conversion of a training job description to the arguments that the class constructor expects. After attaching, if the training job has a Complete status, it can be deploy() ed to create a SageMaker Endpoint and return a Predictor.

If the training job is in progress, attach will block and display log messages from the training job, until the training job completes.

Parameters:
  • training_job_name (str) – The name of the training job to attach to.
  • sagemaker_session (sagemaker.session.Session) – Session object which manages interactions with Amazon SageMaker APIs and any other AWS services needed. If not specified, the estimator creates one using the default AWS configuration chain.
  • model_channel_name (str) – Name of the channel where pre-trained model data will be downloaded (default: ‘model’). If no channel with the same name exists in the training job, this option will be ignored.

Examples

>>> my_estimator.fit(wait=False)
>>> training_job_name = my_estimator.latest_training_job.name
Later on:
>>> attached_estimator = Estimator.attach(training_job_name)
>>> attached_estimator.deploy()
Returns:Instance of the calling Estimator Class with the attached training job.
deploy(initial_instance_count, instance_type, accelerator_type=None, endpoint_name=None, use_compiled_model=False, update_endpoint=False, **kwargs)

Deploy the trained model to an Amazon SageMaker endpoint and return a sagemaker.RealTimePredictor object.

More information: http://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-training.html

Parameters:
  • initial_instance_count (int) – Minimum number of EC2 instances to deploy to an endpoint for prediction.
  • instance_type (str) – Type of EC2 instance to deploy to an endpoint for prediction, for example, ‘ml.c4.xlarge’.
  • accelerator_type (str) – Type of Elastic Inference accelerator to attach to an endpoint for model loading and inference, for example, ‘ml.eia1.medium’. If not specified, no Elastic Inference accelerator will be attached to the endpoint. For more information: https://docs.aws.amazon.com/sagemaker/latest/dg/ei.html
  • endpoint_name (str) – Name to use for creating an Amazon SageMaker endpoint. If not specified, the name of the training job is used.
  • use_compiled_model (bool) – Flag to select whether to use compiled (optimized) model. Default: False.
  • update_endpoint (bool) – Flag to update the model in an existing Amazon SageMaker endpoint. If True, this will deploy a new EndpointConfig to an already existing endpoint and delete resources corresponding to the previous EndpointConfig. Default: False
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A predictor that provides a predict() method,

which can be used to send requests to the Amazon SageMaker endpoint and obtain inferences.

Return type:

sagemaker.predictor.RealTimePredictor

model_data

str – The model location in S3. Only set if Estimator has been fit().

create_model(**kwargs)

Create a SageMaker Model object that can be deployed to an Endpoint.

Parameters:**kwargs – Keyword arguments used by the implemented method for creating the Model.
Returns:A SageMaker Model object. See Model() for full details.
Return type:sagemaker.model.Model
delete_endpoint()

Delete an Amazon SageMaker Endpoint.

Raises:ValueError – If the endpoint does not exist.
transformer(instance_count, instance_type, strategy=None, assemble_with=None, output_path=None, output_kms_key=None, accept=None, env=None, max_concurrent_transforms=None, max_payload=None, tags=None, role=None, volume_kms_key=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 the Estimator.

Parameters:
  • 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’.
  • 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, results are stored to a default bucket.
  • output_kms_key (str) – Optional. KMS key ID for encrypting the transform output (default: None).
  • accept (str) – The content type accepted by the endpoint deployed during the transform job.
  • env (dict) – Environment variables to be set for use during the transform job (default: None).
  • 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 container in MB.
  • tags (list[dict]) – List of tags for labeling a transform job. If none specified, then the tags used for the training job are used for the transform job.
  • role (str) – The ExecutionRoleArn IAM Role ARN for the Model, which is also used during transform jobs. If not specified, the role from the Estimator will be used.
  • volume_kms_key (str) – Optional. KMS key ID for encrypting the volume attached to the ML compute instance (default: None).
training_job_analytics

Return a TrainingJobAnalytics object for the current training job.

get_vpc_config(vpc_config_override='VPC_CONFIG_DEFAULT')

Returns VpcConfig dict either from this Estimator’s subnets and security groups, or else validate and return an optional override value.

class sagemaker.estimator.Estimator(image_name, role, train_instance_count, train_instance_type, train_volume_size=30, train_volume_kms_key=None, train_max_run=86400, input_mode='File', output_path=None, output_kms_key=None, base_job_name=None, sagemaker_session=None, hyperparameters=None, tags=None, subnets=None, security_group_ids=None, model_uri=None, model_channel_name='model', metric_definitions=None, encrypt_inter_container_traffic=False)

Bases: sagemaker.estimator.EstimatorBase

A generic Estimator to train using any supplied algorithm. This class is designed for use with algorithms that don’t have their own, custom class.

Initialize an Estimator instance.

Parameters:
  • image_name (str) – The container image to use for training.
  • 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 artifacts. After the endpoint is created, the inference code might use the IAM role, if it needs to access an AWS resource.
  • train_instance_count (int) – Number of Amazon EC2 instances to use for training.
  • train_instance_type (str) – Type of EC2 instance to use for training, for example, ‘ml.c4.xlarge’.
  • train_volume_size (int) – Size in GB of the EBS volume to use for storing input data during training (default: 30). Must be large enough to store training data if File Mode is used (which is the default).
  • train_volume_kms_key (str) – Optional. KMS key ID for encrypting EBS volume attached to the training instance (default: None).
  • train_max_run (int) – Timeout in seconds for training (default: 24 * 60 * 60). After this amount of time Amazon SageMaker terminates the job regardless of its current status.
  • input_mode (str) –

    The input mode that the algorithm supports (default: ‘File’). Valid modes:

    • ’File’ - Amazon SageMaker copies the training dataset from the S3 location to a local directory.
    • ’Pipe’ - Amazon SageMaker streams data directly from S3 to the container via a Unix-named pipe.

    This argument can be overriden on a per-channel basis using sagemaker.session.s3_input.input_mode.

  • output_path (str) – S3 location for saving the trainig result (model artifacts and output files). If not specified, results are stored to a default bucket. If the bucket with the specific name does not exist, the estimator creates the bucket during the fit() method execution.
  • output_kms_key (str) – Optional. KMS key ID for encrypting the training output (default: None).
  • base_job_name (str) – Prefix for training job name when the fit() method launches. If not specified, the estimator generates a default job name, based on the training image name and current timestamp.
  • sagemaker_session (sagemaker.session.Session) – Session object which manages interactions with Amazon SageMaker APIs and any other AWS services needed. If not specified, the estimator creates one using the default AWS configuration chain.
  • hyperparameters (dict) – Dictionary containing the hyperparameters to initialize this estimator with.
  • tags (list[dict]) – List of tags for labeling a training job. For more, see https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
  • subnets (list[str]) – List of subnet ids. If not specified training job will be created without VPC config.
  • security_group_ids (list[str]) – List of security group ids. If not specified training job will be created without VPC config.
  • model_uri (str) –

    URI where a pre-trained model is stored, either locally or in S3 (default: None). If specified, the estimator will create a channel pointing to the model so the training job can download it. This model can be a ‘model.tar.gz’ from a previous training job, or other artifacts coming from a different source.

    In local mode, this should point to the path in which the model is located and not the file itself, as local Docker containers will try to mount the URI as a volume.

    More information: https://docs.aws.amazon.com/sagemaker/latest/dg/cdf-training.html#td-deserialization

  • model_channel_name (str) – Name of the channel where ‘model_uri’ will be downloaded (default: ‘model’).
  • metric_definitions (list[dict]) – A list of dictionaries that defines the metric(s) used to evaluate the training jobs. Each dictionary contains two keys: ‘Name’ for the name of the metric, and ‘Regex’ for the regular expression used to extract the metric from the logs. This should be defined only for jobs that don’t use an Amazon algorithm.
  • encrypt_inter_container_traffic (bool) – Specifies whether traffic between training containers is encrypted for the training job (default: False).
train_image()

Returns the docker image to use for training.

The fit() method, that does the model training, calls this method to find the image to use for model training.

set_hyperparameters(**kwargs)
hyperparameters()

Returns the hyperparameters as a dictionary to use for training.

The fit() method, that does the model training, calls this method to find the hyperparameters you specified.

create_model(role=None, image=None, predictor_cls=None, serializer=None, deserializer=None, content_type=None, accept=None, vpc_config_override='VPC_CONFIG_DEFAULT', **kwargs)

Create a model to deploy.

Parameters:
  • role (str) – The ExecutionRoleArn IAM Role ARN for the Model, which is also used during transform jobs. If not specified, the role from the Estimator will be used.
  • image (str) – An container image to use for deploying the model. Defaults to the image used for training.
  • predictor_cls (RealTimePredictor) – The predictor class to use when deploying the model.
  • serializer (callable) – Should accept a single argument, the input data, and return a sequence of bytes. May provide a content_type attribute that defines the endpoint request content type
  • deserializer (callable) – Should accept two arguments, the result data and the response content type, and return a sequence of bytes. May provide a content_type attribute that defines th endpoint response Accept content type.
  • content_type (str) – The invocation ContentType, overriding any content_type from the serializer
  • accept (str) – The invocation Accept, overriding any accept from the deserializer.
  • vpc_config_override (dict[str, list[str]]) – Optional override for 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.
  • serializer, deserializer, content_type, and accept arguments are only used to define a default (The) –
  • They are ignored if an explicit predictor class is passed in. Other arguments (RealTimePredictor.) –
  • passed through to the Model class. (are) –

Returns: a Model ready for deployment.

classmethod attach(training_job_name, sagemaker_session=None, model_channel_name='model')

Attach to an existing training job.

Create an Estimator bound to an existing training job, each subclass is responsible to implement _prepare_init_params_from_job_description() as this method delegates the actual conversion of a training job description to the arguments that the class constructor expects. After attaching, if the training job has a Complete status, it can be deploy() ed to create a SageMaker Endpoint and return a Predictor.

If the training job is in progress, attach will block and display log messages from the training job, until the training job completes.

Parameters:
  • training_job_name (str) – The name of the training job to attach to.
  • sagemaker_session (sagemaker.session.Session) – Session object which manages interactions with Amazon SageMaker APIs and any other AWS services needed. If not specified, the estimator creates one using the default AWS configuration chain.
  • model_channel_name (str) – Name of the channel where pre-trained model data will be downloaded (default: ‘model’). If no channel with the same name exists in the training job, this option will be ignored.

Examples

>>> my_estimator.fit(wait=False)
>>> training_job_name = my_estimator.latest_training_job.name
Later on:
>>> attached_estimator = Estimator.attach(training_job_name)
>>> attached_estimator.deploy()
Returns:Instance of the calling Estimator Class with the attached training job.
compile_model(target_instance_family, input_shape, output_path, framework=None, framework_version=None, compile_max_run=300, tags=None, **kwargs)

Compile a Neo model using the input model.

Parameters:
  • target_instance_family (str) – Identifies the device that you want to run your model after compilation, for example: ml_c5. Allowed strings are: ml_c5, ml_m5, ml_c4, ml_m4, jetsontx1, jetsontx2, ml_p2, ml_p3, deeplens, rasp3b
  • input_shape (dict) – Specifies the name and shape of the expected inputs for your trained model in json dictionary form, for example: {‘data’:[1,3,1024,1024]}, or {‘var1’: [1,1,28,28], ‘var2’:[1,1,28,28]}
  • output_path (str) – Specifies where to store the compiled model
  • framework (str) – The framework that is used to train the original model. Allowed values: ‘mxnet’, ‘tensorflow’, ‘pytorch’, ‘onnx’, ‘xgboost’
  • framework_version (str) – The version of the framework
  • compile_max_run (int) – Timeout in seconds for compilation (default: 3 * 60). After this amount of time Amazon SageMaker Neo terminates the compilation job regardless of its current status.
  • tags (list[dict]) – List of tags for labeling a compilation job. For more, see https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A SageMaker Model object. See Model() for full details.

Return type:

sagemaker.model.Model

delete_endpoint()

Delete an Amazon SageMaker Endpoint.

Raises:ValueError – If the endpoint does not exist.
deploy(initial_instance_count, instance_type, accelerator_type=None, endpoint_name=None, use_compiled_model=False, update_endpoint=False, **kwargs)

Deploy the trained model to an Amazon SageMaker endpoint and return a sagemaker.RealTimePredictor object.

More information: http://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-training.html

Parameters:
  • initial_instance_count (int) – Minimum number of EC2 instances to deploy to an endpoint for prediction.
  • instance_type (str) – Type of EC2 instance to deploy to an endpoint for prediction, for example, ‘ml.c4.xlarge’.
  • accelerator_type (str) – Type of Elastic Inference accelerator to attach to an endpoint for model loading and inference, for example, ‘ml.eia1.medium’. If not specified, no Elastic Inference accelerator will be attached to the endpoint. For more information: https://docs.aws.amazon.com/sagemaker/latest/dg/ei.html
  • endpoint_name (str) – Name to use for creating an Amazon SageMaker endpoint. If not specified, the name of the training job is used.
  • use_compiled_model (bool) – Flag to select whether to use compiled (optimized) model. Default: False.
  • update_endpoint (bool) – Flag to update the model in an existing Amazon SageMaker endpoint. If True, this will deploy a new EndpointConfig to an already existing endpoint and delete resources corresponding to the previous EndpointConfig. Default: False
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A predictor that provides a predict() method,

which can be used to send requests to the Amazon SageMaker endpoint and obtain inferences.

Return type:

sagemaker.predictor.RealTimePredictor

enable_network_isolation()

Return True if this Estimator will need network isolation to run.

Returns:Whether this Estimator needs network isolation or not.
Return type:bool
fit(inputs=None, wait=True, logs=True, job_name=None)

Train a model using the input training dataset.

The API calls the Amazon SageMaker CreateTrainingJob API to start model training. The API uses configuration you provided to create the estimator and the specified input training data to send the CreatingTrainingJob request to Amazon SageMaker.

This is a synchronous operation. After the model training successfully completes, you can call the deploy() method to host the model using the Amazon SageMaker hosting services.

Parameters:
  • inputs (str or dict or sagemaker.session.s3_input) –

    Information about the training data. This can be one of three types:

    • (str) the S3 location where training data is saved.
    • (dict[str, str] or dict[str, sagemaker.session.s3_input]) If using multiple channels for
      training data, you can specify a dict mapping channel names to strings or s3_input() objects.
    • (sagemaker.session.s3_input) - channel configuration for S3 data sources that can provide
      additional information as well as the path to the training dataset. See sagemaker.session.s3_input() for full details.
  • wait (bool) – Whether the call should wait until the job completes (default: True).
  • logs (bool) – Whether to show the logs produced by the job. Only meaningful when wait is True (default: True).
  • job_name (str) – Training job name. If not specified, the estimator generates a default job name, based on the training image name and current timestamp.
get_vpc_config(vpc_config_override='VPC_CONFIG_DEFAULT')

Returns VpcConfig dict either from this Estimator’s subnets and security groups, or else validate and return an optional override value.

model_data

str – The model location in S3. Only set if Estimator has been fit().

training_job_analytics

Return a TrainingJobAnalytics object for the current training job.

transformer(instance_count, instance_type, strategy=None, assemble_with=None, output_path=None, output_kms_key=None, accept=None, env=None, max_concurrent_transforms=None, max_payload=None, tags=None, role=None, volume_kms_key=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 the Estimator.

Parameters:
  • 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’.
  • 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, results are stored to a default bucket.
  • output_kms_key (str) – Optional. KMS key ID for encrypting the transform output (default: None).
  • accept (str) – The content type accepted by the endpoint deployed during the transform job.
  • env (dict) – Environment variables to be set for use during the transform job (default: None).
  • 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 container in MB.
  • tags (list[dict]) – List of tags for labeling a transform job. If none specified, then the tags used for the training job are used for the transform job.
  • role (str) – The ExecutionRoleArn IAM Role ARN for the Model, which is also used during transform jobs. If not specified, the role from the Estimator will be used.
  • volume_kms_key (str) – Optional. KMS key ID for encrypting the volume attached to the ML compute instance (default: None).
class sagemaker.estimator.Framework(entry_point, source_dir=None, hyperparameters=None, enable_cloudwatch_metrics=False, container_log_level=20, code_location=None, image_name=None, dependencies=None, **kwargs)

Bases: sagemaker.estimator.EstimatorBase

Base class that cannot be instantiated directly.

Subclasses define functionality pertaining to specific ML frameworks, such as training/deployment images and predictor instances.

Base class initializer. Subclasses which override __init__ should invoke super()

Parameters:
  • entry_point (str) – Path (absolute or relative) to the local Python source file which should be executed as the entry point to training. This should be compatible with either Python 2.7 or Python 3.5.
  • source_dir (str) – Path (absolute or relative) to a directory with any other training source code dependencies aside from the entry point file (default: None). Structure within this directory are preserved when training on Amazon SageMaker.
  • dependencies (list[str]) –

    A list of paths to directories (absolute or relative) with any additional libraries that will be exported to the container (default: []). The library folders will be copied to SageMaker in the same folder where the entrypoint is copied. .. rubric:: Example

    The following call >>> Estimator(entry_point=’train.py’, dependencies=[‘my/libs/common’, ‘virtual-env’]) results in the following inside the container:

    >>> $ ls
    
    >>> opt/ml/code
    >>>     |------ train.py
    >>>     |------ common
    >>>     |------ virtual-env
    
  • hyperparameters (dict) – Hyperparameters that will be used for training (default: None). The hyperparameters are made accessible as a dict[str, str] to the training code on SageMaker. For convenience, this accepts other types for keys and values, but str() will be called to convert them before training.
  • enable_cloudwatch_metrics (bool) – [DEPRECATED] Now there are cloudwatch metrics emitted by all SageMaker training jobs. This will be ignored for now and removed in a further release.
  • container_log_level (int) – Log level to use within the container (default: logging.INFO). Valid values are defined in the Python logging module.
  • code_location (str) – The S3 prefix URI where custom code will be uploaded (default: None). The code file uploaded in S3 is ‘code_location/source/sourcedir.tar.gz’. If not specified, the default code location is s3://default_bucket/job-name/. And code file uploaded to S3 is s3://default_bucket/job-name/source/sourcedir.tar.gz
  • image_name (str) – An alternate image name to use instead of the official Sagemaker image for the framework. This is useful to run one of the Sagemaker supported frameworks with an image containing custom dependencies.
  • **kwargs – Additional kwargs passed to the EstimatorBase constructor.
LAUNCH_PS_ENV_NAME = 'sagemaker_parameter_server_enabled'
LAUNCH_MPI_ENV_NAME = 'sagemaker_mpi_enabled'
MPI_NUM_PROCESSES_PER_HOST = 'sagemaker_mpi_num_of_processes_per_host'
MPI_CUSTOM_MPI_OPTIONS = 'sagemaker_mpi_custom_mpi_options'
hyperparameters()

Return the hyperparameters as a dictionary to use for training.

The fit() method, which trains the model, calls this method to find the hyperparameters.

Returns:The hyperparameters.
Return type:dict[str, str]
train_image()

Return the Docker image to use for training.

The fit() method, which does the model training, calls this method to find the image to use for model training.

Returns:The URI of the Docker image.
Return type:str
classmethod attach(training_job_name, sagemaker_session=None, model_channel_name='model')

Attach to an existing training job.

Create an Estimator bound to an existing training job, each subclass is responsible to implement _prepare_init_params_from_job_description() as this method delegates the actual conversion of a training job description to the arguments that the class constructor expects. After attaching, if the training job has a Complete status, it can be deploy() ed to create a SageMaker Endpoint and return a Predictor.

If the training job is in progress, attach will block and display log messages from the training job, until the training job completes.

Parameters:
  • training_job_name (str) – The name of the training job to attach to.
  • sagemaker_session (sagemaker.session.Session) – Session object which manages interactions with Amazon SageMaker APIs and any other AWS services needed. If not specified, the estimator creates one using the default AWS configuration chain.
  • model_channel_name (str) – Name of the channel where pre-trained model data will be downloaded (default: ‘model’). If no channel with the same name exists in the training job, this option will be ignored.

Examples

>>> my_estimator.fit(wait=False)
>>> training_job_name = my_estimator.latest_training_job.name
Later on:
>>> attached_estimator = Estimator.attach(training_job_name)
>>> attached_estimator.deploy()
Returns:Instance of the calling Estimator Class with the attached training job.
transformer(instance_count, instance_type, strategy=None, assemble_with=None, output_path=None, output_kms_key=None, accept=None, env=None, max_concurrent_transforms=None, max_payload=None, tags=None, role=None, model_server_workers=None, volume_kms_key=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 the Estimator.

Parameters:
  • 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’.
  • 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, results are stored to a default bucket.
  • output_kms_key (str) – Optional. KMS key ID for encrypting the transform output (default: None).
  • accept (str) – The content type accepted by the endpoint deployed during the transform job.
  • env (dict) – Environment variables to be set for use during the transform job (default: None).
  • 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 container in MB.
  • tags (list[dict]) – List of tags for labeling a transform job. If none specified, then the tags used for the training job are used for the transform job.
  • role (str) – The ExecutionRoleArn IAM Role ARN for the Model, which is also used during transform jobs. If not specified, the role from the Estimator will be used.
  • model_server_workers (int) – Optional. The number of worker processes used by the inference server. If None, server will use one worker per vCPU.
  • volume_kms_key (str) – Optional. KMS key ID for encrypting the volume attached to the ML compute instance (default: None).
compile_model(target_instance_family, input_shape, output_path, framework=None, framework_version=None, compile_max_run=300, tags=None, **kwargs)

Compile a Neo model using the input model.

Parameters:
  • target_instance_family (str) – Identifies the device that you want to run your model after compilation, for example: ml_c5. Allowed strings are: ml_c5, ml_m5, ml_c4, ml_m4, jetsontx1, jetsontx2, ml_p2, ml_p3, deeplens, rasp3b
  • input_shape (dict) – Specifies the name and shape of the expected inputs for your trained model in json dictionary form, for example: {‘data’:[1,3,1024,1024]}, or {‘var1’: [1,1,28,28], ‘var2’:[1,1,28,28]}
  • output_path (str) – Specifies where to store the compiled model
  • framework (str) – The framework that is used to train the original model. Allowed values: ‘mxnet’, ‘tensorflow’, ‘pytorch’, ‘onnx’, ‘xgboost’
  • framework_version (str) – The version of the framework
  • compile_max_run (int) – Timeout in seconds for compilation (default: 3 * 60). After this amount of time Amazon SageMaker Neo terminates the compilation job regardless of its current status.
  • tags (list[dict]) – List of tags for labeling a compilation job. For more, see https://docs.aws.amazon.com/sagemaker/latest/dg/API_Tag.html.
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A SageMaker Model object. See Model() for full details.

Return type:

sagemaker.model.Model

create_model(**kwargs)

Create a SageMaker Model object that can be deployed to an Endpoint.

Parameters:**kwargs – Keyword arguments used by the implemented method for creating the Model.
Returns:A SageMaker Model object. See Model() for full details.
Return type:sagemaker.model.Model
delete_endpoint()

Delete an Amazon SageMaker Endpoint.

Raises:ValueError – If the endpoint does not exist.
deploy(initial_instance_count, instance_type, accelerator_type=None, endpoint_name=None, use_compiled_model=False, update_endpoint=False, **kwargs)

Deploy the trained model to an Amazon SageMaker endpoint and return a sagemaker.RealTimePredictor object.

More information: http://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-training.html

Parameters:
  • initial_instance_count (int) – Minimum number of EC2 instances to deploy to an endpoint for prediction.
  • instance_type (str) – Type of EC2 instance to deploy to an endpoint for prediction, for example, ‘ml.c4.xlarge’.
  • accelerator_type (str) – Type of Elastic Inference accelerator to attach to an endpoint for model loading and inference, for example, ‘ml.eia1.medium’. If not specified, no Elastic Inference accelerator will be attached to the endpoint. For more information: https://docs.aws.amazon.com/sagemaker/latest/dg/ei.html
  • endpoint_name (str) – Name to use for creating an Amazon SageMaker endpoint. If not specified, the name of the training job is used.
  • use_compiled_model (bool) – Flag to select whether to use compiled (optimized) model. Default: False.
  • update_endpoint (bool) – Flag to update the model in an existing Amazon SageMaker endpoint. If True, this will deploy a new EndpointConfig to an already existing endpoint and delete resources corresponding to the previous EndpointConfig. Default: False
  • **kwargs – Passed to invocation of create_model(). Implementations may customize create_model() to accept **kwargs to customize model creation during deploy. For more, see the implementation docs.
Returns:

A predictor that provides a predict() method,

which can be used to send requests to the Amazon SageMaker endpoint and obtain inferences.

Return type:

sagemaker.predictor.RealTimePredictor

enable_network_isolation()

Return True if this Estimator will need network isolation to run.

Returns:Whether this Estimator needs network isolation or not.
Return type:bool
fit(inputs=None, wait=True, logs=True, job_name=None)

Train a model using the input training dataset.

The API calls the Amazon SageMaker CreateTrainingJob API to start model training. The API uses configuration you provided to create the estimator and the specified input training data to send the CreatingTrainingJob request to Amazon SageMaker.

This is a synchronous operation. After the model training successfully completes, you can call the deploy() method to host the model using the Amazon SageMaker hosting services.

Parameters:
  • inputs (str or dict or sagemaker.session.s3_input) –

    Information about the training data. This can be one of three types:

    • (str) the S3 location where training data is saved.
    • (dict[str, str] or dict[str, sagemaker.session.s3_input]) If using multiple channels for
      training data, you can specify a dict mapping channel names to strings or s3_input() objects.
    • (sagemaker.session.s3_input) - channel configuration for S3 data sources that can provide
      additional information as well as the path to the training dataset. See sagemaker.session.s3_input() for full details.
  • wait (bool) – Whether the call should wait until the job completes (default: True).
  • logs (bool) – Whether to show the logs produced by the job. Only meaningful when wait is True (default: True).
  • job_name (str) – Training job name. If not specified, the estimator generates a default job name, based on the training image name and current timestamp.
get_vpc_config(vpc_config_override='VPC_CONFIG_DEFAULT')

Returns VpcConfig dict either from this Estimator’s subnets and security groups, or else validate and return an optional override value.

model_data

str – The model location in S3. Only set if Estimator has been fit().

training_job_analytics

Return a TrainingJobAnalytics object for the current training job.