R/debugger.R
In DyfanJones/sagemaker-r-common: R6sagemaker lower level api calls
Defines functions
`!=.CollectionConfig`
`==.CollectionConfig`
get_default_profiler_rule
get_rule_container_image_uri
Documented in
get_default_profiler_rule
get_rule_container_image_uri
# NOTE: This code has been modified from AWS Sagemaker Python:
# https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/debugger.py
#' @include r_utils.R
#' @import R6
#' @import sagemaker.core
#' @import sagemaker.debugger
#' @importFrom jsonlite toJSON
DEBUGGER_FLAG = "USE_SMDEBUG"
#' @title Return the Debugger rule image URI for the given AWS Region.
#' @description For a full list of rule image URIs,
#' see `Use Debugger Docker Images for Built-in or Custom Rules
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-docker-images-rules.html}.
#' @param region (str): A string of AWS Region. For example, ``'us-east-1'``.
#' @return str : Formatted image URI for the given AWS Region and the rule container type.
#' @export
get_rule_container_image_uri <- function(region){
return(ImageUris$new()$retrieve("debugger", region))
}
#' @title Return the default built-in profiler rule with a unique name.
#' @return ProfilerRule: The instance of the built-in ProfilerRule.
#' @export
get_default_profiler_rule <- function(){
default_rule = sagemaker.debugger::ProfilerReport$new()
custom_name = sprintf("%s-%s", default_rule$rule_name, as.integer(Sys.time()))
return(ProfilerRule$new()$sagemaker(default_rule, name=custom_name))
}
#' @title The SageMaker Debugger rule base class that cannot be instantiated directly.
#' @description Debugger rule classes inheriting this RuleBase class are
#' :class:`~sagemaker.debugger.Rule` and :class:`~sagemaker.debugger.ProfilerRule`.
#' Do not directly use the rule base class to instantiate a SageMaker Debugger rule.
#' Use the :class:`~sagemaker.debugger.Rule` classmethods for debugging
#' and the :class:`~sagemaker.debugger.ProfilerRule` classmethods for profiling.
#' @export
RuleBase = R6Class("RuleBase",
public = list(
#' @field name
#' (str): The name of the rule.
name = NULL,
#' @field image_uri
#' (str): The image URI to use the rule.
image_uri = NULL,
#' @field instance_type
#' (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
instance_type = NULL,
#' @field container_local_output_path
#' (str): The local path to store the Rule output.
container_local_output_path = NULL,
#' @field s3_output_path
#' (str): The location in S3 to store the output.
s3_output_path = NULL,
#' @field volume_size_in_gb
#' (int): Size in GB of the EBS volume to use for storing data.
volume_size_in_gb = NULL,
#' @field rule_parameters
#' (dict): A dictionary of parameters for the rule.
rule_parameters = NULL,
#' @description Initialize RuleBase class
#' @param name (str): The name of the rule.
#' @param image_uri (str): The image URI to use the rule.
#' @param instance_type (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
#' @param container_local_output_path (str): The local path to store the Rule output.
#' @param s3_output_path (str): The location in S3 to store the output.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume to use for storing data.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
initialize = function(name=NULL,
image_uri=NULL,
instance_type=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
volume_size_in_gb=NULL,
rule_parameters=NULL){
stopifnot(is.null(name) || is.character(name),
is.null(image_uri) || is.character(image_uri),
is.null(instance_type) || is.character(instance_type),
is.null(container_local_output_path) || is.character(container_local_output_path),
is.null(s3_output_path) || is.character(s3_output_path),
is.null(volume_size_in_gb) || is.numeric(volume_size_in_gb),
is.null(rule_parameters) || is.list(rule_parameters))
self$name = name
self$image_uri = image_uri
self$instance_type = instance_type
self$container_local_output_path = container_local_output_path
self$s3_output_path = s3_output_path
self$volume_size_in_gb = volume_size_in_gb
self$rule_parameters = rule_parameters
},
#' @description format class
format = function(){
return(format_class(self))
}
),
private = list(
# @description Create a dictionary of rule parameters.
# @param source (str): Optional. A source file containing a rule to invoke. If provided,
# you must also provide rule_to_invoke. This can either be an S3 uri or
# a local path.
# @param rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
# If provided, you must also provide source.
# @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
# @return dict: A dictionary of rule parameters.
.set_rule_parameters = function(source=NULL,
rule_to_invoke=NULL,
rule_parameters=NULL){
if ((is.null(source) + is.null(rule_to_invoke)) == 1)
ValueError$new("If you provide a source, you must also provide a rule to invoke (and vice versa).")
merged_rule_params = list()
merged_rule_params = c(merged_rule_params, build_dict("source_s3_uri", source))
merged_rule_params = c(merged_rule_params , build_dict("rule_to_invoke", rule_to_invoke))
merged_rule_params = c(merged_rule_params, rule_parameters %||% list())
return(merged_rule_params)
}
),
lock_objects = F
)
#' @title Debug Rule Class
#' @description The SageMaker Debugger Rule class configures *debugging* rules to debug your training job.
#' The debugging rules analyze tensor outputs from your training job
#' and monitor conditions that are critical for the success of the training
#' job.
#' SageMaker Debugger comes pre-packaged with built-in *debugging* rules.
#' For example, the debugging rules can detect whether gradients are getting too large or
#' too small, or if a model is overfitting.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' You can also write your own rules using the custom rule classmethod.
#' @export
Rule = R6Class("Rule",
inherit = RuleBase,
public = list(
#' @field collection_configs
#' A list of :class:`~sagemaker.debugger.CollectionConfig
collection_configs = NULL,
#' @field actions
#' Placeholder
actions = NULL,
#' @description Configure the debugging rules using the following classmethods.
#' @param name (str): The name of the rule.
#' @param image_uri (str): The image URI to use the rule.
#' @param instance_type (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
#' @param container_local_output_path (str): The local path to store the Rule output.
#' @param s3_output_path (str): The location in S3 to store the output.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume to use for storing data.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
#' @param collections_to_save ([sagemaker.debugger.CollectionConfig]): Optional. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved.
#' @param actions :
initialize = function(name=NULL,
image_uri=NULL,
instance_type=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
volume_size_in_gb=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
super$initialize(
name,
image_uri,
instance_type,
container_local_output_path,
s3_output_path,
volume_size_in_gb,
rule_parameters)
self$collection_configs = collections_to_save
self$actions = actions
},
#' @description Initialize a ``Rule`` object for a \code{built-in} debugging rule.
#' @param base_config (dict): Required. This is the base rule config dictionary returned from the
#' :class:\code{sagemaker.debugger} method.
#' For example, \code{sagemaker.debugger::dead_relu()}.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' @param name (str): Optional. The name of the debugger rule. If one is not provided,
#' the name of the base_config will be used.
#' @param container_local_output_path (str): Optional. The local path in the rule processing
#' container.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path for debugging data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param other_trials_s3_input_paths ([str]): Optional. The Amazon S3 input paths
#' of other trials to use the SimilarAcrossRuns rule.
#' @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
#' @param collections_to_save (:class:\code{sagemaker.debugger::CollectionConfig}):
#' Optional. A list
#' of :class:\code{sagemaker.debugger::CollectionConfig} objects to be saved.
#' @param actions :
#' @return :class:`~sagemaker.debugger.Rule`: An instance of the built-in rule.
sagemaker = function(base_config,
name=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
other_trials_s3_input_paths=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
merged_rule_params = list()
if (!is.null(rule_parameters) && !islistempty(rule_parameters$rule_to_invoke))
RuntimeError$new(
"You cannot provide a 'rule_to_invoke' for SageMaker rules.",
"Either remove the rule_to_invoke or use a custom rule."
)
if (!is.null(actions) && !inherits(actions, c("Action", "ActionList")))
RuntimeError$new("`actions` must be of type `Action` or `ActionList`!")
if (!is.null(other_trials_s3_input_paths)){
for (index in seq_along(other_trials_s3_input_paths)){
merged_rule_params[[sprintf("other_trial_%s",index-1)]] = other_trials_s3_input_paths[index]
}
}
default_rule_params = base_config[["DebugRuleConfiguration"]][["RuleParameters"]] %||% list()
merged_rule_params = c(merged_rule_params, default_rule_params)
merged_rule_params = c(merged_rule_params, rule_parameters %||% list())
base_config_collections = list()
for (config in (base_config[["CollectionConfigurations"]] %||% list())){
collection_name = NULL
collection_parameters = list()
for (i in seq_along(config)){
key = names(config)[i]
value = config[[i]]
if (key == "CollectionName")
collection_name = value
if (key == "CollectionParameters")
collection_parameters = value
}
base_config_collections = c(
base_config_collections, CollectionConfig$new(name=collection_name, parameters=collection_parameters))
}
self$initialize(
name=name %||% base_config[["DebugRuleConfiguration"]][["RuleConfigurationName"]],
image_uri="DEFAULT_RULE_EVALUATOR_IMAGE",
instance_type=NULL,
container_local_output_path=container_local_output_path,
s3_output_path=s3_output_path,
volume_size_in_gb=NULL,
rule_parameters=merged_rule_params,
collections_to_save=collections_to_save %||% base_config_collections,
actions=actions
)
return(self)
},
#' @description Initialize a ``Rule`` object for a *custom* debugging rule.
#' You can create a custom rule that analyzes tensors emitted
#' during the training of a model
#' and monitors conditions that are critical for the success of a training
#' job. For more information, see `Create Debugger Custom Rules for Training Job Analysis`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-custom-rules.html}.
#' @param name (str): Required. The name of the debugger rule.
#' @param image_uri (str): Required. The URI of the image to be used by the debugger rule.
#' @param instance_type (str): Required. Type of EC2 instance to use, for example,
#' 'ml.c4.xlarge'.
#' @param volume_size_in_gb (int): Required. Size in GB of the EBS volume
#' to use for storing data.
#' @param source (str): Optional. A source file containing a rule to invoke. If provided,
#' you must also provide rule_to_invoke. This can either be an S3 uri or
#' a local path.
#' @param rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
#' If provided, you must also provide source.
#' @param container_local_output_path (str): Optional. The local path in the container.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path for debugging data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param other_trials_s3_input_paths ([str]): Optional. The Amazon S3 input paths
#' of other trials to use the SimilarAcrossRuns rule.
#' @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
#' @param collections_to_save ([sagemaker.debugger.CollectionConfig]): Optional. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved.
#' @param actions :
#' @return :class:`~sagemaker.debugger.Rule`: The instance of the custom rule.
custom = function(name,
image_uri,
instance_type,
volume_size_in_gb,
source=NULL,
rule_to_invoke=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
other_trials_s3_input_paths=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
if(!is.null(actions) && !(inherits(actions, "Action") || inherits(actions, "ActionList")))
RuntimeError$new("`actions` must be of type `Action` or `ActionList`!")
merged_rule_params = private$.set_rule_parameters(
source, rule_to_invoke, other_trials_s3_input_paths, rule_parameters)
self$initialize(
name=name,
image_uri=image_uri,
instance_type=instance_type,
container_local_output_path=container_local_output_path,
s3_output_path=s3_output_path,
volume_size_in_gb=volume_size_in_gb,
rule_parameters=merged_rule_params,
collections_to_save=collections_to_save %||% list(),
actions=actions)
return(self)
},
#' @description Prepare actions for Debugger Rule.
#' @param training_job_name (str): The training job name. To be set as the default training job
#' prefix for the StopTraining action if it is specified.
prepare_actions = function(training_job_name){
if (is.null(self$actions)){
# user cannot manually specify action_json in rule_parameters for actions.
self$rule_parameters$action_json <- NULL
return(NULL)}
self$actions$update_training_job_prefix_if_not_specified(training_job_name)
action_params = list("action_json"= self$actions$serialize())
self$rule_parameters = modifyList(self$rule_parameters, action_params)
},
#' @description Generates a request dictionary using the parameters provided when initializing object.
#' @return dict: An portion of an API request as a dictionary.
to_debugger_rule_config_list = function(){
debugger_rule_config_request = list(
"RuleConfigurationName"= self$name,
"RuleEvaluatorImage"= self$image_uri)
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("InstanceType", self$instance_type))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("VolumeSizeInGB", self$volume_size_in_gb))
debugger_rule_config_request = c(debugger_rule_config_request,
build_dict("LocalPath", self$container_local_output_path))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("S3OutputPath", self$s3_output_path))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("RuleParameters", self$rule_parameters))
return(debugger_rule_config_request)
}
),
private = list(
# Set rule parameters for Debugger Rule.
# Args:
# source (str): Optional. A source file containing a rule to invoke. If provided,
# you must also provide rule_to_invoke. This can either be an S3 uri or
# a local path.
# rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
# If provided, you must also provide source.
# other_trials_s3_input_paths ([str]): Optional. S3 input paths for other trials.
# rule_parameters (dict): Optional. A dictionary of parameters for the rule.
# Returns:
# dict: A dictionary of rule parameters.
.set_rule_parameters = function(source,
rule_to_invoke,
other_trials_s3_input_paths,
rule_parameters){
merged_rule_params = list()
if (!is.null(other_trials_s3_input_paths)){
for (index in seq_along(other_trials_s3_input_paths)){
merged_rule_params[[sprintf("other_trial_%s",index-1)]] = other_trials_s3_input_paths[index]
}
}
merged_rule_params = c(
merged_rule_params,
super$.set_rule_parameters(source, rule_to_invoke, rule_parameters))
return(merged_rule_params)
}
)
)
#' @title The SageMaker Debugger ProfilerRule class configures *profiling* rules.
#' @description SageMaker Debugger profiling rules automatically analyze
#' hardware system resource utilization and framework metrics of a
#' training job to identify performance bottlenecks.
#' SageMaker Debugger comes pre-packaged with built-in *profiling* rules.
#' For example, the profiling rules can detect if GPUs are underutilized due to CPU bottlenecks or
#' IO bottlenecks.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' You can also write your own profiling rules using the Amazon SageMaker
#' Debugger APIs.
#' @export
ProfilerRule = R6Class("ProfilerRule",
inherit = RuleBase,
public = list(
#' @description Initialize a ``ProfilerRule`` object for a *built-in* profiling rule.
#' The rule analyzes system and framework metrics of a given
#' training job to identify performance bottlenecks.
#' @param base_config (\code{sagemaker.debugger::ProfilerReport}): The base rule configuration object
#' returned from the \code{sagemaker.debugger} method.
#' For example, \code{sagemaker.debugger::ProfilerReport$new()}.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' @param name (str): The name of the profiler rule. If one is not provided,
#' the name of the base_config will be used.
#' @param container_local_output_path (str): The path in the container.
#' @param s3_output_path (str): The location in Amazon S3 to store the profiling output data.
#' The default Debugger output path for profiling data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/profiler-output/.
#' @return :class:`~sagemaker.debugger.ProfilerRule`:
#' The instance of the built-in ProfilerRule.
sagemaker = function(base_config,
name=NULL,
container_local_output_path=NULL,
s3_output_path=NULL){
self$initialize(
name = name %||% base_config$rule_name,
image_uri = "DEFAULT_RULE_EVALUATOR_IMAGE",
instance_type = NULL,
container_local_output_path = container_local_output_path,
s3_output_path = s3_output_path,
volume_size_in_gb = NULL,
rule_parameters = base_config$rule_parameters
)
return(self)
},
#' @description Initialize a ``ProfilerRule`` object for a *custom* profiling rule.
#' You can create a rule that
#' analyzes system and framework metrics emitted during the training of a model and
#' monitors conditions that are critical for the success of a
#' training job.
#' @param name (str): The name of the profiler rule.
#' @param image_uri (str): The URI of the image to be used by the proflier rule.
#' @param instance_type (str): Type of EC2 instance to use, for example,
#' 'ml.c4.xlarge'.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume
#' to use for storing data.
#' @param source (str): A source file containing a rule to invoke. If provided,
#' you must also provide rule_to_invoke. This can either be an S3 uri or
#' a local path.
#' @param rule_to_invoke (str): The name of the rule to invoke within the source.
#' If provided, you must also provide the source.
#' @param container_local_output_path (str): The path in the container.
#' @param s3_output_path (str): The location in Amazon S3 to store the output.
#' The default Debugger output path for profiling data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/profiler-output/.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
#' @return :class:`~sagemaker.debugger.ProfilerRule`:
#' The instance of the custom ProfilerRule.
custom = function(name,
image_uri,
instance_type,
volume_size_in_gb,
source=NULL,
rule_to_invoke=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
rule_parameters=NULL){
merged_rule_params = super$.set_rule_parameters(source, rule_to_invoke, rule_parameters)
self$initialize(
name = name,
image_uri = image_uri,
instance_type = instance_type,
container_local_output_path = container_local_output_path,
s3_output_path = s3_output_path,
volume_size_in_gb = volume_size_in_gb,
rule_parameters = merged_rule_params
)
return(self)
},
#' @description Generates a request dictionary using the parameters provided when initializing object.
#' @return lict: An portion of an API request as a dictionary.
to_profiler_rule_config_list = function(){
profiler_rule_config_request = list(
"RuleConfigurationName"= self$name,
"RuleEvaluatorImage"= self$image_uri
)
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("InstanceType", self$instance_type))
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("VolumeSizeInGB", self$volume_size_in_gb))
profiler_rule_config_request = c(profiler_rule_config_request,
build_dict("LocalPath", self$container_local_output_path))
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("S3OutputPath", self$s3_output_path))
if (!islistempty(self$rule_parameters)){
profiler_rule_config_request[["RuleParameters"]] = self$rule_parameters
for (i in seq_along(profiler_rule_config_request[["RuleParameters"]])){
k = names(profiler_rule_config_request[["RuleParameters"]])[i]
v = profiler_rule_config_request[["RuleParameters"]][[i]]
profiler_rule_config_request[["RuleParameters"]][[k]] = to_str(v)
}
}
return(profiler_rule_config_request)
}
),
lock_objects = F
)
#' @title Create a Debugger hook configuration object to save the tensor for debugging.
#' @description \code{DebuggerHookConfig} provides options to customize how debugging
#' information is emitted and saved. This high-level \code{DebuggerHookConfig} class
#' runs based on the `smdebug.SaveConfig
#' \url{https://github.com/awslabs/sagemaker-debugger/blob/master/docs/api.md#saveconfig}
#' class.
#' @export
DebuggerHookConfig = R6Class("DebuggerHookConfig",
public = list(
#' @field s3_output_path
#' The location in Amazon S3 to store the output tensors
s3_output_path = NULL,
#' @field container_local_output_path
#' The local path in the container
container_local_output_path = NULL,
#' @field hook_parameters
#' A dictionary of parameters
hook_parameters = NULL,
#' @field collection_configs
#' A list of :class:`~sagemaker.debugger.CollectionConfig
collection_configs = NULL,
#' @description Initialize the DebuggerHookConfig instance.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param container_local_output_path (str): Optional. The local path in the container.
#' @param hook_parameters (dict): Optional. A dictionary of parameters.
#' @param collection_configs ([sagemaker.debugger.CollectionConfig]): Required. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved
#' at the \code{s3_output_path}.
initialize = function(s3_output_path = NULL,
container_local_output_path = NULL,
hook_parameters = NULL,
collection_configs = NULL){
self$s3_output_path = s3_output_path
self$container_local_output_path = container_local_output_path
self$hook_parameters = hook_parameters
self$collection_configs = collection_configs
},
#' @description Generate a request dictionary using the parameters when initializing the object.
#' @return dict: An portion of an API request as a dictionary.
to_request_list = function(){
debugger_hook_config_request = list("S3OutputPath"= self$s3_output_path)
debugger_hook_config_request[["LocalPath"]] = self$container_local_output_path
debugger_hook_config_request[["HookParameters"]] = self$hook_parameters
if(!is.null(self$collection_configs))
debugger_hook_config_request[[
"CollectionConfigurations"
]] = lapply(self$collection_configs, function(collection_config)
collection_config$to_request_list()
)
return(debugger_hook_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
#' @title TensorBoardOutputConfig Class
#' @description Create a tensor ouput configuration object for debugging visualizations on TensorBoard
#' @export
TensorBoardOutputConfig = R6Class("TensorBoardOutputConfig",
public = list(
#' @field s3_output_path
#' The location in Amazon S3 to store the output.
s3_output_path = NULL,
#' @field container_local_output_path
#' The local path in the container.
container_local_output_path = NULL,
#' @description Initialize the TensorBoardOutputConfig instance.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output.
#' @param container_local_output_path (str): Optional. The local path in the container.
initialize = function(s3_output_path,
container_local_output_path=NULL){
self$s3_output_path = s3_output_path
self$container_local_output_path = container_local_output_path
},
#' @description Generate a request dictionary using the instances attributes.
#' @return dict: An portion of an API request as a dictionary.
to_request_list = function(){
tensorboard_output_config_request = list("S3OutputPath"= self$s3_output_path)
tensorboard_output_config_request$LocalPath = self$container_local_output_path
return(tensorboard_output_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
#' @title CollectionConfig Class
#' @description Creates tensor collections for SageMaker Debugger
#' @export
CollectionConfig = R6Class("CollectionConfig",
public = list(
#' @field name
#' The name of the collection configuration.
name = NULL,
#' @field parameters
#' The parameters for the collection configuration.
parameters = NULL,
#' @description Constructor for collection configuration.
#' @param name (str): Required. The name of the collection configuration.
#' @param parameters (dict): Optional. The parameters for the collection
#' configuration.
initialize = function(name,
parameters=NULL){
self$name = name
self$parameters = parameters
},
#' @description Generate a request dictionary using the parameters initializing the object.
#' @return dict: A portion of an API request as a dictionary.
to_request_list = function(){
collection_config_request = list("CollectionName"= self$name)
collection_config_request[["CollectionParameters"]] = self$parameters
return(collection_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
# Equal method override.
# Args:
# other: Object to test equality against.
#' @export
`==.CollectionConfig` <- function(self, other){
if (!inherits(other, "CollectionConfig"))
TypeError$new(
"CollectionConfig is only comparable with other CollectionConfig objects."
)
eq_name = self$name == other$name
# added an extra step for rebustness of check
eq_parm = (
if(is.null(self$parameters) || is.null(other$parameters))
identical(self$parameters, other$self$parameters)
else
identical(
self$parameters[order(names(self$parameters))],
other$parameters[order(names(other$parameters))])
)
return(eq_name && eq_parm)
}
# Not-equal method override.
# Args:
# other: Object to test equality against.
#' @export
`!=.CollectionConfig` <- function(self, other){
if (!inherits(other, "CollectionConfig"))
TypeError$new(
"CollectionConfig is only comparable with other CollectionConfig objects."
)
no_eq_name = self$name != other$name
# added an extra step for rebustness of check
no_eq_parm = (
if(is.null(self$parameters) || is.null(other$parameters))
!identical(self$parameters, other$self$parameters)
else
!identical(
self$parameters[order(names(self$parameters))],
other$parameters[order(names(other$parameters))])
)
return(no_eq_name || no_eq_parm)
}
# NOTE: Not a 100% how to implement hash method in R6. Possibly not required in R.
#
# def __hash__(self):
# """Hash method override."""
# return hash((self.name, tuple(sorted((self.parameters or {}).items()))))
DyfanJones/sagemaker-r-common documentation built on June 14, 2022, 10:31 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions `!=.CollectionConfig` `==.CollectionConfig` get_default_profiler_rule get_rule_container_image_uri
Documented in get_default_profiler_rule get_rule_container_image_uri
# NOTE: This code has been modified from AWS Sagemaker Python:
# https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/debugger.py
#' @include r_utils.R
#' @import R6
#' @import sagemaker.core
#' @import sagemaker.debugger
#' @importFrom jsonlite toJSON
DEBUGGER_FLAG = "USE_SMDEBUG"
#' @title Return the Debugger rule image URI for the given AWS Region.
#' @description For a full list of rule image URIs,
#' see `Use Debugger Docker Images for Built-in or Custom Rules
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-docker-images-rules.html}.
#' @param region (str): A string of AWS Region. For example, ``'us-east-1'``.
#' @return str : Formatted image URI for the given AWS Region and the rule container type.
#' @export
get_rule_container_image_uri <- function(region){
return(ImageUris$new()$retrieve("debugger", region))
}
#' @title Return the default built-in profiler rule with a unique name.
#' @return ProfilerRule: The instance of the built-in ProfilerRule.
#' @export
get_default_profiler_rule <- function(){
default_rule = sagemaker.debugger::ProfilerReport$new()
custom_name = sprintf("%s-%s", default_rule$rule_name, as.integer(Sys.time()))
return(ProfilerRule$new()$sagemaker(default_rule, name=custom_name))
}
#' @title The SageMaker Debugger rule base class that cannot be instantiated directly.
#' @description Debugger rule classes inheriting this RuleBase class are
#' :class:`~sagemaker.debugger.Rule` and :class:`~sagemaker.debugger.ProfilerRule`.
#' Do not directly use the rule base class to instantiate a SageMaker Debugger rule.
#' Use the :class:`~sagemaker.debugger.Rule` classmethods for debugging
#' and the :class:`~sagemaker.debugger.ProfilerRule` classmethods for profiling.
#' @export
RuleBase = R6Class("RuleBase",
public = list(
#' @field name
#' (str): The name of the rule.
name = NULL,
#' @field image_uri
#' (str): The image URI to use the rule.
image_uri = NULL,
#' @field instance_type
#' (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
instance_type = NULL,
#' @field container_local_output_path
#' (str): The local path to store the Rule output.
container_local_output_path = NULL,
#' @field s3_output_path
#' (str): The location in S3 to store the output.
s3_output_path = NULL,
#' @field volume_size_in_gb
#' (int): Size in GB of the EBS volume to use for storing data.
volume_size_in_gb = NULL,
#' @field rule_parameters
#' (dict): A dictionary of parameters for the rule.
rule_parameters = NULL,
#' @description Initialize RuleBase class
#' @param name (str): The name of the rule.
#' @param image_uri (str): The image URI to use the rule.
#' @param instance_type (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
#' @param container_local_output_path (str): The local path to store the Rule output.
#' @param s3_output_path (str): The location in S3 to store the output.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume to use for storing data.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
initialize = function(name=NULL,
image_uri=NULL,
instance_type=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
volume_size_in_gb=NULL,
rule_parameters=NULL){
stopifnot(is.null(name) || is.character(name),
is.null(image_uri) || is.character(image_uri),
is.null(instance_type) || is.character(instance_type),
is.null(container_local_output_path) || is.character(container_local_output_path),
is.null(s3_output_path) || is.character(s3_output_path),
is.null(volume_size_in_gb) || is.numeric(volume_size_in_gb),
is.null(rule_parameters) || is.list(rule_parameters))
self$name = name
self$image_uri = image_uri
self$instance_type = instance_type
self$container_local_output_path = container_local_output_path
self$s3_output_path = s3_output_path
self$volume_size_in_gb = volume_size_in_gb
self$rule_parameters = rule_parameters
},
#' @description format class
format = function(){
return(format_class(self))
}
),
private = list(
# @description Create a dictionary of rule parameters.
# @param source (str): Optional. A source file containing a rule to invoke. If provided,
# you must also provide rule_to_invoke. This can either be an S3 uri or
# a local path.
# @param rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
# If provided, you must also provide source.
# @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
# @return dict: A dictionary of rule parameters.
.set_rule_parameters = function(source=NULL,
rule_to_invoke=NULL,
rule_parameters=NULL){
if ((is.null(source) + is.null(rule_to_invoke)) == 1)
ValueError$new("If you provide a source, you must also provide a rule to invoke (and vice versa).")
merged_rule_params = list()
merged_rule_params = c(merged_rule_params, build_dict("source_s3_uri", source))
merged_rule_params = c(merged_rule_params , build_dict("rule_to_invoke", rule_to_invoke))
merged_rule_params = c(merged_rule_params, rule_parameters %||% list())
return(merged_rule_params)
}
),
lock_objects = F
)
#' @title Debug Rule Class
#' @description The SageMaker Debugger Rule class configures *debugging* rules to debug your training job.
#' The debugging rules analyze tensor outputs from your training job
#' and monitor conditions that are critical for the success of the training
#' job.
#' SageMaker Debugger comes pre-packaged with built-in *debugging* rules.
#' For example, the debugging rules can detect whether gradients are getting too large or
#' too small, or if a model is overfitting.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' You can also write your own rules using the custom rule classmethod.
#' @export
Rule = R6Class("Rule",
inherit = RuleBase,
public = list(
#' @field collection_configs
#' A list of :class:`~sagemaker.debugger.CollectionConfig
collection_configs = NULL,
#' @field actions
#' Placeholder
actions = NULL,
#' @description Configure the debugging rules using the following classmethods.
#' @param name (str): The name of the rule.
#' @param image_uri (str): The image URI to use the rule.
#' @param instance_type (str): Type of EC2 instance to use. For example, 'ml.c4.xlarge'.
#' @param container_local_output_path (str): The local path to store the Rule output.
#' @param s3_output_path (str): The location in S3 to store the output.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume to use for storing data.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
#' @param collections_to_save ([sagemaker.debugger.CollectionConfig]): Optional. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved.
#' @param actions :
initialize = function(name=NULL,
image_uri=NULL,
instance_type=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
volume_size_in_gb=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
super$initialize(
name,
image_uri,
instance_type,
container_local_output_path,
s3_output_path,
volume_size_in_gb,
rule_parameters)
self$collection_configs = collections_to_save
self$actions = actions
},
#' @description Initialize a ``Rule`` object for a \code{built-in} debugging rule.
#' @param base_config (dict): Required. This is the base rule config dictionary returned from the
#' :class:\code{sagemaker.debugger} method.
#' For example, \code{sagemaker.debugger::dead_relu()}.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' @param name (str): Optional. The name of the debugger rule. If one is not provided,
#' the name of the base_config will be used.
#' @param container_local_output_path (str): Optional. The local path in the rule processing
#' container.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path for debugging data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param other_trials_s3_input_paths ([str]): Optional. The Amazon S3 input paths
#' of other trials to use the SimilarAcrossRuns rule.
#' @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
#' @param collections_to_save (:class:\code{sagemaker.debugger::CollectionConfig}):
#' Optional. A list
#' of :class:\code{sagemaker.debugger::CollectionConfig} objects to be saved.
#' @param actions :
#' @return :class:`~sagemaker.debugger.Rule`: An instance of the built-in rule.
sagemaker = function(base_config,
name=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
other_trials_s3_input_paths=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
merged_rule_params = list()
if (!is.null(rule_parameters) && !islistempty(rule_parameters$rule_to_invoke))
RuntimeError$new(
"You cannot provide a 'rule_to_invoke' for SageMaker rules.",
"Either remove the rule_to_invoke or use a custom rule."
)
if (!is.null(actions) && !inherits(actions, c("Action", "ActionList")))
RuntimeError$new("`actions` must be of type `Action` or `ActionList`!")
if (!is.null(other_trials_s3_input_paths)){
for (index in seq_along(other_trials_s3_input_paths)){
merged_rule_params[[sprintf("other_trial_%s",index-1)]] = other_trials_s3_input_paths[index]
}
}
default_rule_params = base_config[["DebugRuleConfiguration"]][["RuleParameters"]] %||% list()
merged_rule_params = c(merged_rule_params, default_rule_params)
merged_rule_params = c(merged_rule_params, rule_parameters %||% list())
base_config_collections = list()
for (config in (base_config[["CollectionConfigurations"]] %||% list())){
collection_name = NULL
collection_parameters = list()
for (i in seq_along(config)){
key = names(config)[i]
value = config[[i]]
if (key == "CollectionName")
collection_name = value
if (key == "CollectionParameters")
collection_parameters = value
}
base_config_collections = c(
base_config_collections, CollectionConfig$new(name=collection_name, parameters=collection_parameters))
}
self$initialize(
name=name %||% base_config[["DebugRuleConfiguration"]][["RuleConfigurationName"]],
image_uri="DEFAULT_RULE_EVALUATOR_IMAGE",
instance_type=NULL,
container_local_output_path=container_local_output_path,
s3_output_path=s3_output_path,
volume_size_in_gb=NULL,
rule_parameters=merged_rule_params,
collections_to_save=collections_to_save %||% base_config_collections,
actions=actions
)
return(self)
},
#' @description Initialize a ``Rule`` object for a *custom* debugging rule.
#' You can create a custom rule that analyzes tensors emitted
#' during the training of a model
#' and monitors conditions that are critical for the success of a training
#' job. For more information, see `Create Debugger Custom Rules for Training Job Analysis`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-custom-rules.html}.
#' @param name (str): Required. The name of the debugger rule.
#' @param image_uri (str): Required. The URI of the image to be used by the debugger rule.
#' @param instance_type (str): Required. Type of EC2 instance to use, for example,
#' 'ml.c4.xlarge'.
#' @param volume_size_in_gb (int): Required. Size in GB of the EBS volume
#' to use for storing data.
#' @param source (str): Optional. A source file containing a rule to invoke. If provided,
#' you must also provide rule_to_invoke. This can either be an S3 uri or
#' a local path.
#' @param rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
#' If provided, you must also provide source.
#' @param container_local_output_path (str): Optional. The local path in the container.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path for debugging data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param other_trials_s3_input_paths ([str]): Optional. The Amazon S3 input paths
#' of other trials to use the SimilarAcrossRuns rule.
#' @param rule_parameters (dict): Optional. A dictionary of parameters for the rule.
#' @param collections_to_save ([sagemaker.debugger.CollectionConfig]): Optional. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved.
#' @param actions :
#' @return :class:`~sagemaker.debugger.Rule`: The instance of the custom rule.
custom = function(name,
image_uri,
instance_type,
volume_size_in_gb,
source=NULL,
rule_to_invoke=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
other_trials_s3_input_paths=NULL,
rule_parameters=NULL,
collections_to_save=NULL,
actions=NULL){
if(!is.null(actions) && !(inherits(actions, "Action") || inherits(actions, "ActionList")))
RuntimeError$new("`actions` must be of type `Action` or `ActionList`!")
merged_rule_params = private$.set_rule_parameters(
source, rule_to_invoke, other_trials_s3_input_paths, rule_parameters)
self$initialize(
name=name,
image_uri=image_uri,
instance_type=instance_type,
container_local_output_path=container_local_output_path,
s3_output_path=s3_output_path,
volume_size_in_gb=volume_size_in_gb,
rule_parameters=merged_rule_params,
collections_to_save=collections_to_save %||% list(),
actions=actions)
return(self)
},
#' @description Prepare actions for Debugger Rule.
#' @param training_job_name (str): The training job name. To be set as the default training job
#' prefix for the StopTraining action if it is specified.
prepare_actions = function(training_job_name){
if (is.null(self$actions)){
# user cannot manually specify action_json in rule_parameters for actions.
self$rule_parameters$action_json <- NULL
return(NULL)}
self$actions$update_training_job_prefix_if_not_specified(training_job_name)
action_params = list("action_json"= self$actions$serialize())
self$rule_parameters = modifyList(self$rule_parameters, action_params)
},
#' @description Generates a request dictionary using the parameters provided when initializing object.
#' @return dict: An portion of an API request as a dictionary.
to_debugger_rule_config_list = function(){
debugger_rule_config_request = list(
"RuleConfigurationName"= self$name,
"RuleEvaluatorImage"= self$image_uri)
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("InstanceType", self$instance_type))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("VolumeSizeInGB", self$volume_size_in_gb))
debugger_rule_config_request = c(debugger_rule_config_request,
build_dict("LocalPath", self$container_local_output_path))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("S3OutputPath", self$s3_output_path))
debugger_rule_config_request = c(debugger_rule_config_request, build_dict("RuleParameters", self$rule_parameters))
return(debugger_rule_config_request)
}
),
private = list(
# Set rule parameters for Debugger Rule.
# Args:
# source (str): Optional. A source file containing a rule to invoke. If provided,
# you must also provide rule_to_invoke. This can either be an S3 uri or
# a local path.
# rule_to_invoke (str): Optional. The name of the rule to invoke within the source.
# If provided, you must also provide source.
# other_trials_s3_input_paths ([str]): Optional. S3 input paths for other trials.
# rule_parameters (dict): Optional. A dictionary of parameters for the rule.
# Returns:
# dict: A dictionary of rule parameters.
.set_rule_parameters = function(source,
rule_to_invoke,
other_trials_s3_input_paths,
rule_parameters){
merged_rule_params = list()
if (!is.null(other_trials_s3_input_paths)){
for (index in seq_along(other_trials_s3_input_paths)){
merged_rule_params[[sprintf("other_trial_%s",index-1)]] = other_trials_s3_input_paths[index]
}
}
merged_rule_params = c(
merged_rule_params,
super$.set_rule_parameters(source, rule_to_invoke, rule_parameters))
return(merged_rule_params)
}
)
)
#' @title The SageMaker Debugger ProfilerRule class configures *profiling* rules.
#' @description SageMaker Debugger profiling rules automatically analyze
#' hardware system resource utilization and framework metrics of a
#' training job to identify performance bottlenecks.
#' SageMaker Debugger comes pre-packaged with built-in *profiling* rules.
#' For example, the profiling rules can detect if GPUs are underutilized due to CPU bottlenecks or
#' IO bottlenecks.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' You can also write your own profiling rules using the Amazon SageMaker
#' Debugger APIs.
#' @export
ProfilerRule = R6Class("ProfilerRule",
inherit = RuleBase,
public = list(
#' @description Initialize a ``ProfilerRule`` object for a *built-in* profiling rule.
#' The rule analyzes system and framework metrics of a given
#' training job to identify performance bottlenecks.
#' @param base_config (\code{sagemaker.debugger::ProfilerReport}): The base rule configuration object
#' returned from the \code{sagemaker.debugger} method.
#' For example, \code{sagemaker.debugger::ProfilerReport$new()}.
#' For a full list of built-in rules for debugging, see
#' `List of Debugger Built-in Rules`
#' \url{https://docs.aws.amazon.com/sagemaker/latest/dg/debugger-built-in-rules.html}.
#' @param name (str): The name of the profiler rule. If one is not provided,
#' the name of the base_config will be used.
#' @param container_local_output_path (str): The path in the container.
#' @param s3_output_path (str): The location in Amazon S3 to store the profiling output data.
#' The default Debugger output path for profiling data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/profiler-output/.
#' @return :class:`~sagemaker.debugger.ProfilerRule`:
#' The instance of the built-in ProfilerRule.
sagemaker = function(base_config,
name=NULL,
container_local_output_path=NULL,
s3_output_path=NULL){
self$initialize(
name = name %||% base_config$rule_name,
image_uri = "DEFAULT_RULE_EVALUATOR_IMAGE",
instance_type = NULL,
container_local_output_path = container_local_output_path,
s3_output_path = s3_output_path,
volume_size_in_gb = NULL,
rule_parameters = base_config$rule_parameters
)
return(self)
},
#' @description Initialize a ``ProfilerRule`` object for a *custom* profiling rule.
#' You can create a rule that
#' analyzes system and framework metrics emitted during the training of a model and
#' monitors conditions that are critical for the success of a
#' training job.
#' @param name (str): The name of the profiler rule.
#' @param image_uri (str): The URI of the image to be used by the proflier rule.
#' @param instance_type (str): Type of EC2 instance to use, for example,
#' 'ml.c4.xlarge'.
#' @param volume_size_in_gb (int): Size in GB of the EBS volume
#' to use for storing data.
#' @param source (str): A source file containing a rule to invoke. If provided,
#' you must also provide rule_to_invoke. This can either be an S3 uri or
#' a local path.
#' @param rule_to_invoke (str): The name of the rule to invoke within the source.
#' If provided, you must also provide the source.
#' @param container_local_output_path (str): The path in the container.
#' @param s3_output_path (str): The location in Amazon S3 to store the output.
#' The default Debugger output path for profiling data is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/profiler-output/.
#' @param rule_parameters (dict): A dictionary of parameters for the rule.
#' @return :class:`~sagemaker.debugger.ProfilerRule`:
#' The instance of the custom ProfilerRule.
custom = function(name,
image_uri,
instance_type,
volume_size_in_gb,
source=NULL,
rule_to_invoke=NULL,
container_local_output_path=NULL,
s3_output_path=NULL,
rule_parameters=NULL){
merged_rule_params = super$.set_rule_parameters(source, rule_to_invoke, rule_parameters)
self$initialize(
name = name,
image_uri = image_uri,
instance_type = instance_type,
container_local_output_path = container_local_output_path,
s3_output_path = s3_output_path,
volume_size_in_gb = volume_size_in_gb,
rule_parameters = merged_rule_params
)
return(self)
},
#' @description Generates a request dictionary using the parameters provided when initializing object.
#' @return lict: An portion of an API request as a dictionary.
to_profiler_rule_config_list = function(){
profiler_rule_config_request = list(
"RuleConfigurationName"= self$name,
"RuleEvaluatorImage"= self$image_uri
)
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("InstanceType", self$instance_type))
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("VolumeSizeInGB", self$volume_size_in_gb))
profiler_rule_config_request = c(profiler_rule_config_request,
build_dict("LocalPath", self$container_local_output_path))
profiler_rule_config_request = c(profiler_rule_config_request, build_dict("S3OutputPath", self$s3_output_path))
if (!islistempty(self$rule_parameters)){
profiler_rule_config_request[["RuleParameters"]] = self$rule_parameters
for (i in seq_along(profiler_rule_config_request[["RuleParameters"]])){
k = names(profiler_rule_config_request[["RuleParameters"]])[i]
v = profiler_rule_config_request[["RuleParameters"]][[i]]
profiler_rule_config_request[["RuleParameters"]][[k]] = to_str(v)
}
}
return(profiler_rule_config_request)
}
),
lock_objects = F
)
#' @title Create a Debugger hook configuration object to save the tensor for debugging.
#' @description \code{DebuggerHookConfig} provides options to customize how debugging
#' information is emitted and saved. This high-level \code{DebuggerHookConfig} class
#' runs based on the `smdebug.SaveConfig
#' \url{https://github.com/awslabs/sagemaker-debugger/blob/master/docs/api.md#saveconfig}
#' class.
#' @export
DebuggerHookConfig = R6Class("DebuggerHookConfig",
public = list(
#' @field s3_output_path
#' The location in Amazon S3 to store the output tensors
s3_output_path = NULL,
#' @field container_local_output_path
#' The local path in the container
container_local_output_path = NULL,
#' @field hook_parameters
#' A dictionary of parameters
hook_parameters = NULL,
#' @field collection_configs
#' A list of :class:`~sagemaker.debugger.CollectionConfig
collection_configs = NULL,
#' @description Initialize the DebuggerHookConfig instance.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output tensors.
#' The default Debugger output path is created under the
#' default output path of the :class:`~sagemaker.estimator.Estimator` class.
#' For example,
#' s3://sagemaker-<region>-<12digit_account_id>/<training-job-name>/debug-output/.
#' @param container_local_output_path (str): Optional. The local path in the container.
#' @param hook_parameters (dict): Optional. A dictionary of parameters.
#' @param collection_configs ([sagemaker.debugger.CollectionConfig]): Required. A list
#' of :class:`~sagemaker.debugger.CollectionConfig` objects to be saved
#' at the \code{s3_output_path}.
initialize = function(s3_output_path = NULL,
container_local_output_path = NULL,
hook_parameters = NULL,
collection_configs = NULL){
self$s3_output_path = s3_output_path
self$container_local_output_path = container_local_output_path
self$hook_parameters = hook_parameters
self$collection_configs = collection_configs
},
#' @description Generate a request dictionary using the parameters when initializing the object.
#' @return dict: An portion of an API request as a dictionary.
to_request_list = function(){
debugger_hook_config_request = list("S3OutputPath"= self$s3_output_path)
debugger_hook_config_request[["LocalPath"]] = self$container_local_output_path
debugger_hook_config_request[["HookParameters"]] = self$hook_parameters
if(!is.null(self$collection_configs))
debugger_hook_config_request[[
"CollectionConfigurations"
]] = lapply(self$collection_configs, function(collection_config)
collection_config$to_request_list()
)
return(debugger_hook_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
#' @title TensorBoardOutputConfig Class
#' @description Create a tensor ouput configuration object for debugging visualizations on TensorBoard
#' @export
TensorBoardOutputConfig = R6Class("TensorBoardOutputConfig",
public = list(
#' @field s3_output_path
#' The location in Amazon S3 to store the output.
s3_output_path = NULL,
#' @field container_local_output_path
#' The local path in the container.
container_local_output_path = NULL,
#' @description Initialize the TensorBoardOutputConfig instance.
#' @param s3_output_path (str): Optional. The location in Amazon S3 to store the output.
#' @param container_local_output_path (str): Optional. The local path in the container.
initialize = function(s3_output_path,
container_local_output_path=NULL){
self$s3_output_path = s3_output_path
self$container_local_output_path = container_local_output_path
},
#' @description Generate a request dictionary using the instances attributes.
#' @return dict: An portion of an API request as a dictionary.
to_request_list = function(){
tensorboard_output_config_request = list("S3OutputPath"= self$s3_output_path)
tensorboard_output_config_request$LocalPath = self$container_local_output_path
return(tensorboard_output_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
#' @title CollectionConfig Class
#' @description Creates tensor collections for SageMaker Debugger
#' @export
CollectionConfig = R6Class("CollectionConfig",
public = list(
#' @field name
#' The name of the collection configuration.
name = NULL,
#' @field parameters
#' The parameters for the collection configuration.
parameters = NULL,
#' @description Constructor for collection configuration.
#' @param name (str): Required. The name of the collection configuration.
#' @param parameters (dict): Optional. The parameters for the collection
#' configuration.
initialize = function(name,
parameters=NULL){
self$name = name
self$parameters = parameters
},
#' @description Generate a request dictionary using the parameters initializing the object.
#' @return dict: A portion of an API request as a dictionary.
to_request_list = function(){
collection_config_request = list("CollectionName"= self$name)
collection_config_request[["CollectionParameters"]] = self$parameters
return(collection_config_request)
},
#' @description format class
format = function(){
return(format_class(self))
}
)
)
# Equal method override.
# Args:
# other: Object to test equality against.
#' @export
`==.CollectionConfig` <- function(self, other){
if (!inherits(other, "CollectionConfig"))
TypeError$new(
"CollectionConfig is only comparable with other CollectionConfig objects."
)
eq_name = self$name == other$name
# added an extra step for rebustness of check
eq_parm = (
if(is.null(self$parameters) || is.null(other$parameters))
identical(self$parameters, other$self$parameters)
else
identical(
self$parameters[order(names(self$parameters))],
other$parameters[order(names(other$parameters))])
)
return(eq_name && eq_parm)
}
# Not-equal method override.
# Args:
# other: Object to test equality against.
#' @export
`!=.CollectionConfig` <- function(self, other){
if (!inherits(other, "CollectionConfig"))
TypeError$new(
"CollectionConfig is only comparable with other CollectionConfig objects."
)
no_eq_name = self$name != other$name
# added an extra step for rebustness of check
no_eq_parm = (
if(is.null(self$parameters) || is.null(other$parameters))
!identical(self$parameters, other$self$parameters)
else
!identical(
self$parameters[order(names(self$parameters))],
other$parameters[order(names(other$parameters))])
)
return(no_eq_name || no_eq_parm)
}
# NOTE: Not a 100% how to implement hash method in R6. Possibly not required in R.
#
# def __hash__(self):
# """Hash method override."""
# return hash((self.name, tuple(sorted((self.parameters or {}).items()))))
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.