R/amazon_object2vec.R
In DyfanJones/sagemaker-r-mlframework: sagemaker machine learning developed by amazon
# NOTE: This code has been modified from AWS Sagemaker Python:
# https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/amazon/object2vec.py
#' @include r_utils.R
#' @import R6
#' @import sagemaker.common
#' @import sagemaker.mlcore
#' @import lgr
#' @title A general-purpose neural embedding algorithm that is highly customizable.
#' @description It can learn low-dimensional dense embeddings of high-dimensional objects. The embeddings
#' are learned in a way that preserves the semantics of the relationship between pairs of
#' objects in the original space in the embedding space.
#' @export
Object2Vec = R6Class("Object2Vec",
inherit = sagemaker.mlcore::AmazonAlgorithmEstimatorBase,
public = list(
#' @field repo_name
#' sagemaker repo name for framework
repo_name = "object2vec",
#' @field repo_version
#' version of framework
repo_version = 1,
#' @field MINI_BATCH_SIZE
#' The size of each mini-batch to use when training.
MINI_BATCH_SIZE = 32,
#' @field .module
#' mimic python module
.module = "sagemaker.amazon.object2vec",
#' @description Object2Vec is :class:`Estimator` used for anomaly detection.
#' This Estimator may be fit via calls to
#' :meth:`~sagemaker.amazon.amazon_estimator.AmazonAlgorithmEstimatorBase.fit`.
#' There is an utility
#' :meth:`~sagemaker.amazon.amazon_estimator.AmazonAlgorithmEstimatorBase.record_set`
#' that can be used to upload data to S3 and creates
#' :class:`~sagemaker.amazon.amazon_estimator.RecordSet` to be passed to
#' the `fit` call.
#' After this Estimator is fit, model data is stored in S3. The model
#' may be deployed to an Amazon SageMaker Endpoint by invoking
#' :meth:`~sagemaker.amazon.estimator.EstimatorBase.deploy`. As well as
#' deploying an Endpoint, deploy returns a
#' :class:`~sagemaker.amazon.Predictor` object that can be used for
#' inference calls using the trained model hosted in the SageMaker
#' Endpoint.
#' Object2Vec Estimators can be configured by setting hyperparameters.
#' The available hyperparameters for Object2Vec are documented below.
#' For further information on the AWS Object2Vec algorithm, please
#' consult AWS technical documentation:
#' https://docs.aws.amazon.com/sagemaker/latest/dg/object2vec.html
#' @param 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 accessing AWS resource.
#' @param instance_count (int): Number of Amazon EC2 instances to use
#' for training.
#' @param instance_type (str): Type of EC2 instance to use for training,
#' for example, 'ml.c4.xlarge'.
#' @param epochs (int): Total number of epochs for SGD training
#' @param enc0_max_seq_len (int): Maximum sequence length
#' @param enc0_vocab_size (int): Vocabulary size of tokens
#' @param enc_dim (int): Optional. Dimension of the output of the embedding
#' layer
#' @param mini_batch_size (int): Optional. mini batch size for SGD training
#' @param early_stopping_patience (int): Optional. The allowed number of
#' consecutive epochs without improvement before early stopping is
#' applied
#' @param early_stopping_tolerance (float): Optional. The value used to
#' determine whether the algorithm has made improvement between two
#' consecutive epochs for early stopping
#' @param dropout (float): Optional. Dropout probability on network layers
#' @param weight_decay (float): Optional. Weight decay parameter during
#' optimization
#' @param bucket_width (int): Optional. The allowed difference between data
#' sequence length when bucketing is enabled
#' @param num_classes (int): Optional. Number of classes for classification
#' @param training (ignored for regression problems)
#' @param mlp_layers (int): Optional. Number of MLP layers in the network
#' @param mlp_dim (int): Optional. Dimension of the output of MLP layer
#' @param mlp_activation (str): Optional. Type of activation function for the
#' MLP layer
#' @param output_layer (str): Optional. Type of output layer
#' @param optimizer (str): Optional. Type of optimizer for training
#' @param learning_rate (float): Optional. Learning rate for SGD training
#' @param negative_sampling_rate (int): Optional. Negative sampling rate
#' @param comparator_list (str): Optional. Customization of comparator
#' operator
#' @param tied_token_embedding_weight (bool): Optional. Tying of token
#' embedding layer weight
#' @param token_embedding_storage_type (str): Optional. Type of token
#' embedding storage
#' @param enc0_network (str): Optional. Network model of encoder "enc0"
#' @param enc1_network (str): Optional. Network model of encoder "enc1"
#' @param enc0_cnn_filter_width (int): Optional. CNN filter width
#' @param enc1_cnn_filter_width (int): Optional. CNN filter width
#' @param enc1_max_seq_len (int): Optional. Maximum sequence length
#' @param enc0_token_embedding_dim (int): Optional. Output dimension of token
#' embedding layer
#' @param enc1_token_embedding_dim (int): Optional. Output dimension of token
#' embedding layer
#' @param enc1_vocab_size (int): Optional. Vocabulary size of tokens
#' @param enc0_layers (int): Optional. Number of layers in encoder
#' @param enc1_layers (int): Optional. Number of layers in encoder
#' @param enc0_freeze_pretrained_embedding (bool): Optional. Freeze pretrained
#' embedding weights
#' @param enc1_freeze_pretrained_embedding (bool): Optional. Freeze pretrained
#' embedding weights
#' @param ... : base class keyword argument values.
initialize = function(role,
instance_count,
instance_type,
epochs,
enc0_max_seq_len,
enc0_vocab_size,
enc_dim=NULL,
mini_batch_size=NULL,
early_stopping_patience=NULL,
early_stopping_tolerance=NULL,
dropout=NULL,
weight_decay=NULL,
bucket_width=NULL,
num_classes=NULL,
mlp_layers=NULL,
mlp_dim=NULL,
mlp_activation=NULL,
output_layer=NULL,
optimizer=NULL,
learning_rate=NULL,
negative_sampling_rate=NULL,
comparator_list=NULL,
tied_token_embedding_weight=NULL,
token_embedding_storage_type=NULL,
enc0_network=NULL,
enc1_network=NULL,
enc0_cnn_filter_width=NULL,
enc1_cnn_filter_width=NULL,
enc1_max_seq_len=NULL,
enc0_token_embedding_dim=NULL,
enc1_token_embedding_dim=NULL,
enc1_vocab_size=NULL,
enc0_layers=NULL,
enc1_layers=NULL,
enc0_freeze_pretrained_embedding=NULL,
enc1_freeze_pretrained_embedding=NULL,
...){
private$.enc_dim = Hyperparameter$new("enc_dim", list(Validation$new()$ge(4), Validation$new()$le(10000)), "An integer in [4, 10000]", DataTypes$new()$int, obj = self)
private$.mini_batch_size = Hyperparameter$new("mini_batch_size", list(Validation$new()$ge(1), Validation$new()$le(10000)), "An integer in [1, 10000]", DataTypes$new()$int, obj = self)
private$.epochs = Hyperparameter$new("epochs", list(Validation$new()$ge(1), Validation$new()$le(100)), "An integer in [1, 100]", DataTypes$new()$int, obj = self)
private$.early_stopping_patience = Hyperparameter$new(
"early_stopping_patience", list(Validation$new()$ge(1), Validation$new()$le(5)), "An integer in [1, 5]", DataTypes$new()$int, obj = self
)
private$.early_stopping_tolerance = Hyperparameter$new(
"early_stopping_tolerance", list(Validation$new()$ge(1e-06), Validation$new()$le(0.1)), "A float in [1e-06, 0.1]", DataTypes$new()$float, obj = self
)
private$.dropout = Hyperparameter$new("dropout", list(Validation$new()$ge(0.0), Validation$new()$le(1.0)), "A float in [0.0, 1.0]", DataTypes$new()$float, obj = self)
private$.weight_decay = Hyperparameter$new("weight_decay", list(Validation$new()$ge(0.0), Validation$new()$le(10000.0)), "A float in [0.0, 10000.0]", DataTypes$new()$float, obj = self)
private$.bucket_width = Hyperparameter$new("bucket_width", list(Validation$new()$ge(0), Validation$new()$le(100)), "An integer in [0, 100]", DataTypes$new()$int, obj = self)
private$.num_classes = Hyperparameter$new("num_classes", list(Validation$new()$ge(2), Validation$new()$le(30)), "An integer in [2, 30]", DataTypes$new()$int, obj = self)
private$.mlp_layers = Hyperparameter$new("mlp_layers", list(Validation$new()$ge(1), Validation$new()$le(10)), "An integer in [1, 10]", DataTypes$new()$int, obj = self)
private$.mlp_dim = Hyperparameter$new("mlp_dim", list(Validation$new()$ge(2), Validation$new()$le(10000)), "An integer in [2, 10000]", DataTypes$new()$int, obj = self)
private$.mlp_activation = Hyperparameter$new(
"mlp_activation", Validation$new()$isin(c("tanh", "relu", "linear")), 'One of "tanh", "relu", "linear"', DataTypes$new()$str, obj = self
)
private$.output_layer = Hyperparameter$new(
"output_layer",
Validation$new()$isin(c("softmax", "mean_squared_error")),
'One of "softmax", "mean_squared_error"',
DataTypes$new()$str,
obj = self
)
private$.optimizer = Hyperparameter$new(
"optimizer",
Validation$new()$isin(c("adagrad", "adam", "rmsprop", "sgd", "adadelta")),
'One of "adagrad", "adam", "rmsprop", "sgd", "adadelta"',
DataTypes$new()$str,
obj = self
)
private$.learning_rate = Hyperparameter$new("learning_rate", list(Validation$new()$ge(1e-06), Validation$new()$le(1.0)), "A float in [1e-06, 1.0]", DataTypes$new()$float, obj = self)
private$.negative_sampling_rate = Hyperparameter$new(
"negative_sampling_rate", list(Validation$new()$ge(0), Validation$new()$le(100)), "An integer in [0, 100]", DataTypes$new()$int, obj = self
)
private$.comparator_list = Hyperparameter$new(
"comparator_list",
private$.list_check_subset(c("hadamard", "concat", "abs_diff")),
'Comma-separated of hadamard, concat, abs_diff. E.g. "hadamard,abs_diff"',
DataTypes$new()$str,
obj = self
)
private$.tied_token_embedding_weight = Hyperparameter$new(
"tied_token_embedding_weight", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
private$.token_embedding_storage_type = Hyperparameter$new(
"token_embedding_storage_type",
Validation$new()$isin(c("dense", "row_sparse")),
'One of "dense", "row_sparse"',
DataTypes$new()$str,
obj = self
)
private$.enc0_network = Hyperparameter$new(
"enc0_network",
Validation$new()$isin(c("hcnn", "bilstm", "pooled_embedding")),
'One of "hcnn", "bilstm", "pooled_embedding"',
DataTypes$new()$str,
obj = self
)
private$.enc1_network = Hyperparameter$new(
"enc1_network",
Validation$new()$isin(c("hcnn", "bilstm", "pooled_embedding", "enc0")),
'One of "hcnn", "bilstm", "pooled_embedding", "enc0"',
DataTypes$new()$str,
obj = self
)
private$.enc0_cnn_filter_width = Hyperparameter$new("enc0_cnn_filter_width", list(Validation$new()$ge(1), Validation$new()$le(9)), "An integer in [1, 9]", DataTypes$new()$int, obj = self)
private$.enc1_cnn_filter_width = Hyperparameter$new("enc1_cnn_filter_width", list(Validation$new()$ge(1), Validation$new()$le(9)), "An integer in [1, 9]", DataTypes$new()$int, obj = self)
private$.enc0_max_seq_len = Hyperparameter$new("enc0_max_seq_len", list(Validation$new()$ge(1), Validation$new()$le(5000)), "An integer in [1, 5000]", DataTypes$new()$int, obj = self)
private$.enc1_max_seq_len = Hyperparameter$new("enc1_max_seq_len", list(Validation$new()$ge(1), Validation$new()$le(5000)), "An integer in [1, 5000]", DataTypes$new()$int, obj = self)
private$.enc0_token_embedding_dim = Hyperparameter$new(
"enc0_token_embedding_dim", list(Validation$new()$ge(2), Validation$new()$le(1000)), "An integer in [2, 1000]", DataTypes$new()$int, obj = self
)
private$.enc1_token_embedding_dim = Hyperparameter$new(
"enc1_token_embedding_dim", list(Validation$new()$ge(2), Validation$new()$le(1000)), "An integer in [2, 1000]", DataTypes$new()$int, obj = self
)
private$.enc0_vocab_size = Hyperparameter$new("enc0_vocab_size", list(Validation$new()$ge(2), Validation$new()$le(3000000)), "An integer in [2, 3000000]", DataTypes$new()$int, obj = self)
private$.enc1_vocab_size = Hyperparameter$new("enc1_vocab_size", list(Validation$new()$ge(2), Validation$new()$le(3000000)), "An integer in [2, 3000000]", DataTypes$new()$int, obj = self)
private$.enc0_layers = Hyperparameter$new("enc0_layers", list(Validation$new()$ge(1), Validation$new()$le(4)), "An integer in [1, 4]", DataTypes$new()$int, obj = self)
private$.enc1_layers = Hyperparameter$new("enc1_layers", list(Validation$new()$ge(1), Validation$new()$le(4)), "An integer in [1, 4]", DataTypes$new()$int, obj = self)
private$.enc0_freeze_pretrained_embedding = Hyperparameter$new(
"enc0_freeze_pretrained_embedding", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
private$.enc1_freeze_pretrained_embedding = Hyperparameter$new(
"enc1_freeze_pretrained_embedding", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
super$initialize(role, instance_count, instance_type, ...)
self$enc_dim = enc_dim
self$mini_batch_size = mini_batch_size
self$epochs = epochs
self$early_stopping_patience = early_stopping_patience
self$early_stopping_tolerance = early_stopping_tolerance
self$dropout = dropout
self$weight_decay = weight_decay
self$bucket_width = bucket_width
self$num_classes = num_classes
self$mlp_layers = mlp_layers
self$mlp_dim = mlp_dim
self$mlp_activation = mlp_activation
self$output_layer = output_layer
self$optimizer = optimizer
self$learning_rate = learning_rate
self$negative_sampling_rate = negative_sampling_rate
self$comparator_list = comparator_list
self$tied_token_embedding_weight = tied_token_embedding_weight
self$token_embedding_storage_type = token_embedding_storage_type
self$enc0_network = enc0_network
self$enc1_network = enc1_network
self$enc0_cnn_filter_width = enc0_cnn_filter_width
self$enc1_cnn_filter_width = enc1_cnn_filter_width
self$enc0_max_seq_len = enc0_max_seq_len
self$enc1_max_seq_len = enc1_max_seq_len
self$enc0_token_embedding_dim = enc0_token_embedding_dim
self$enc1_token_embedding_dim = enc1_token_embedding_dim
self$enc0_vocab_size = enc0_vocab_size
self$enc1_vocab_size = enc1_vocab_size
self$enc0_layers = enc0_layers
self$enc1_layers = enc1_layers
self$enc0_freeze_pretrained_embedding = enc0_freeze_pretrained_embedding
self$enc1_freeze_pretrained_embedding = enc1_freeze_pretrained_embedding
},
#' @description Return a :class:`~sagemaker.amazon.Object2VecModel` referencing the
#' latest s3 model data produced by this Estimator.
#' @param vpc_config_override (dict[str, list[str]]): Optional override for VpcConfig set on
#' the model. Default: use subnets and security groups from this Estimator.
#' * 'Subnets' (list[str]): List of subnet ids.
#' * 'SecurityGroupIds' (list[str]): List of security group ids.
#' @param ... : Additional kwargs passed to the Object2VecModel constructor.
create_model = function(vpc_config_override="VPC_CONFIG_DEFAULT",
...){
return(Object2VecModel$new(
self$model_data,
self$role,
sagemaker_session=self$sagemaker_session,
vpc_config=self$get_vpc_config(vpc_config_override),
...
)
)
},
#' @description Set hyperparameters needed for training. This method will also
#' validate ``source_dir``.
#' @param records (RecordSet) – The records to train this Estimator on.
#' @param mini_batch_size (int or None) – The size of each mini-batch to use
#' when training. If None, a default value will be used.
#' @param job_name (str): Name of the training job to be created. If not
#' specified, one is generated, using the base name given to the
#' constructor if applicable.
.prepare_for_training = function(records,
mini_batch_size=NULL,
job_name=NULL){
if (is.null(mini_batch_size))
mini_batch_size = self$MINI_BATCH_SIZE
super$.prepare_for_training(
records, mini_batch_size=mini_batch_size, job_name=job_name
)
}
),
private = list(
# --------- User Active binding to mimic Python's Descriptor Class ---------
.epochs=NULL,
.enc_dim=NULL,
.mini_batch_size=NULL,
.early_stopping_patience=NULL,
.early_stopping_tolerance=NULL,
.dropout=NULL,
.weight_decay=NULL,
.bucket_width=NULL,
.num_classes=NULL,
.mlp_layers=NULL,
.mlp_dim=NULL,
.mlp_activation=NULL,
.output_layer=NULL,
.optimizer=NULL,
.learning_rate=NULL,
.negative_sampling_rate=NULL,
.comparator_list=NULL,
.tied_token_embedding_weight=NULL,
.token_embedding_storage_type=NULL,
.enc0_network=NULL,
.enc1_network=NULL,
.enc0_cnn_filter_width=NULL,
.enc1_cnn_filter_width=NULL,
.enc0_max_seq_len=NULL,
.enc1_max_seq_len=NULL,
.enc0_token_embedding_dim=NULL,
.enc1_token_embedding_dim=NULL,
.enc0_vocab_size=NULL,
.enc1_vocab_size=NULL,
.enc0_layers=NULL,
.enc1_layers=NULL,
.enc0_freeze_pretrained_embedding=NULL,
.enc1_freeze_pretrained_embedding=NULL,
# ---------
.list_check_subset = function(valid_super_list){
valid_superset = unique(valid_super_list)
validate = function(value){
if(!inherits(value, "character"))
return(FALSE)
val_list = lapply(split_str(value), trimws)
return(all(val_list %in% valid_superset))
}
return(validate)
}
),
active = list(
# --------- User Active binding to mimic Python's Descriptor Class ---------
#' @field epochs
#' Total number of epochs for SGD training
epochs = function(value){
if(missing(value))
return(private$.epochs$descriptor)
private$.epochs$descriptor = value
},
#' @field enc_dim
#' Dimension of the output of the embedding layer
enc_dim = function(value){
if(missing(value))
return(private$.enc_dim$descriptor)
private$.enc_dim$descriptor = value
},
#' @field mini_batch_size
#' mini batch size for SGD training
mini_batch_size = function(value){
if(missing(value))
return(private$.mini_batch_size$descriptor)
private$.mini_batch_size$descriptor = value
},
#' @field early_stopping_patience
#' The allowed number of consecutive epochs without
#' improvement before early stopping is applied
early_stopping_patience = function(value){
if(missing(value))
return(private$.early_stopping_patience$descriptor)
private$.early_stopping_patience$descriptor = value
},
#' @field early_stopping_tolerance
#' The value used to determine whether the algorithm
#' has made improvement between two consecutive epochs for early stopping
early_stopping_tolerance = function(value){
if(missing(value))
return(private$.early_stopping_tolerance$descriptor)
private$.early_stopping_tolerance$descriptor = value
},
#' @field dropout
#' Dropout probability on network layers
dropout = function(value){
if(missing(value))
return(private$.dropout$descriptor)
private$.dropout$descriptor = value
},
#' @field weight_decay
#' Weight decay parameter during optimization
weight_decay = function(value){
if(missing(value))
return(private$.weight_decay$descriptor)
private$.weight_decay$descriptor = value
},
#' @field bucket_width
#' The allowed difference between data sequence length when bucketing is enabled
bucket_width = function(value){
if(missing(value))
return(private$.bucket_width$descriptor)
private$.bucket_width$descriptor = value
},
#' @field num_classes
#' Number of classes for classification
num_classes = function(value){
if(missing(value))
return(private$.num_classes$descriptor)
private$.num_classes$descriptor = value
},
#' @field mlp_layers
#' Number of MLP layers in the network
mlp_layers = function(value){
if(missing(value))
return(private$.mlp_layers$descriptor)
private$.mlp_layers$descriptor = value
},
#' @field mlp_dim
#' Dimension of the output of MLP layer
mlp_dim = function(value){
if(missing(value))
return(private$.mlp_dim$descriptor)
private$.mlp_dim$descriptor = value
},
#' @field mlp_activation
#' Type of activation function for the MLP layer
mlp_activation = function(value){
if(missing(value))
return(private$.mlp_activation$descriptor)
private$.mlp_activation$descriptor = value
},
#' @field output_layer
#' Type of output layer
output_layer = function(value){
if(missing(value))
return(private$.output_layer$descriptor)
private$.output_layer$descriptor = value
},
#' @field optimizer
#' Type of optimizer for training
optimizer = function(value){
if(missing(value))
return(private$.optimizer$descriptor)
private$.optimizer$descriptor = value
},
#' @field learning_rate
#' Learning rate for SGD training
learning_rate = function(value){
if(missing(value))
return(private$.learning_rate$descriptor)
private$.learning_rate$descriptor = value
},
#' @field negative_sampling_rate
#' Negative sampling rate
negative_sampling_rate = function(value){
if(missing(value))
return(private$.negative_sampling_rate$descriptor)
private$.negative_sampling_rate$descriptor = value
},
#' @field comparator_list
#' Customization of comparator operator
comparator_list = function(value){
if(missing(value))
return(private$.comparator_list$descriptor)
private$.comparator_list$descriptor = value
},
#' @field tied_token_embedding_weight
#' Tying of token embedding layer weight
tied_token_embedding_weight = function(value){
if(missing(value))
return(private$.tied_token_embedding_weight$descriptor)
private$.tied_token_embedding_weight$descriptor = value
},
#' @field token_embedding_storage_type
#' Type of token embedding storage
token_embedding_storage_type = function(value){
if(missing(value))
return(private$.token_embedding_storage_type$descriptor)
private$.token_embedding_storage_type$descriptor = value
},
#' @field enc0_network
#' Network model of encoder "enc0"
enc0_network = function(value){
if(missing(value))
return(private$.enc0_network$descriptor)
private$.enc0_network$descriptor = value
},
#' @field enc1_network
#' Network model of encoder "enc1"
enc1_network = function(value){
if(missing(value))
return(private$.enc1_network$descriptor)
private$.enc1_network$descriptor = value
},
#' @field enc0_cnn_filter_width
#' CNN filter width
enc0_cnn_filter_width = function(value){
if(missing(value))
return(private$.enc0_cnn_filter_width$descriptor)
private$.enc0_cnn_filter_width$descriptor = value
},
#' @field enc1_cnn_filter_width
#' CNN filter width
enc1_cnn_filter_width = function(value){
if(missing(value))
return(private$.enc1_cnn_filter_width$descriptor)
private$.enc1_cnn_filter_width$descriptor = value
},
#' @field enc0_max_seq_len
#' Maximum sequence length
enc0_max_seq_len = function(value){
if(missing(value))
return(private$.enc0_max_seq_len$descriptor)
private$.enc0_max_seq_len$descriptor = value
},
#' @field enc1_max_seq_len
#' Maximum sequence length
enc1_max_seq_len = function(value){
if(missing(value))
return(private$.enc1_max_seq_len$descriptor)
private$.enc1_max_seq_len$descriptor = value
},
#' @field enc0_token_embedding_dim
#' Output dimension of token embedding layer
enc0_token_embedding_dim = function(value){
if(missing(value))
return(private$.enc0_token_embedding_dim$descriptor)
private$.enc0_token_embedding_dim$descriptor = value
},
#' @field enc1_token_embedding_dim
#' Output dimension of token embedding layer
enc1_token_embedding_dim = function(value){
if(missing(value))
return(private$.enc1_token_embedding_dim$descriptor)
private$.enc1_token_embedding_dim$descriptor = value
},
#' @field enc0_vocab_size
#' Vocabulary size of tokens
enc0_vocab_size = function(value){
if(missing(value))
return(private$.enc0_vocab_size$descriptor)
private$.enc0_vocab_size$descriptor = value
},
#' @field enc1_vocab_size
#' Vocabulary size of tokens
enc1_vocab_size = function(value){
if(missing(value))
return(private$.enc1_vocab_size$descriptor)
private$.enc1_vocab_size$descriptor = value
},
#' @field enc0_layers
#' Number of layers in encoder
enc0_layers = function(value){
if(missing(value))
return(private$.enc0_layers$descriptor)
private$.enc0_layers$descriptor = value
},
#' @field enc1_layers
#' Number of layers in encoder
enc1_layers = function(value){
if(missing(value))
return(private$.enc1_layers$descriptor)
private$.enc1_layers$descriptor = value
},
#' @field enc0_freeze_pretrained_embedding
#' Freeze pretrained embedding weights
enc0_freeze_pretrained_embedding = function(value){
if(missing(value))
return(private$.enc0_freeze_pretrained_embedding$descriptor)
private$.enc0_freeze_pretrained_embedding$descriptor = value
},
#' @field enc1_freeze_pretrained_embedding
#' Freeze pretrained embedding weights
enc1_freeze_pretrained_embedding = function(value){
if(missing(value))
return(private$.enc1_freeze_pretrained_embedding$descriptor)
private$.enc1_freeze_pretrained_embedding$descriptor = value
}
),
lock_objects = F
)
#' @title Reference Object2Vec s3 model data.
#' @description Calling :meth:`~sagemaker.model.Model.deploy` creates an Endpoint and returns a
#' Predictor that calculates anomaly scores for datapoints.
#' @export
Object2VecModel = R6Class("Object2VecModel",
inherit = sagemaker.mlcore::Model,
public = list(
#' @description Initialize Object2VecModel class
#' @param model_data (str): The S3 location of a SageMaker model data
#' ``.tar.gz`` file.
#' @param 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.
#' @param 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.
#' @param ... : Keyword arguments passed to the ``FrameworkModel``
#' initializer.
initialize = function(model_data,
role,
sagemaker_session=NULL,
...){
sagemaker_session = sagemaker_session %||% sagemaker.core::Session$new()
image_uri = sagemaker.core::ImageUris$new()$retrieve(
Object2Vec$public_fields$repo_name,
sagemaker_session$paws_region_name,
version=Object2Vec$public_fields$repo_version
)
super$initialize(
image_uri,
model_data,
role,
predictor_cls=Predictor,
sagemaker_session=sagemaker_session,
...
)
}
),
lock_objects = F
)
DyfanJones/sagemaker-r-mlframework documentation built on March 18, 2022, 7:41 a.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.
# NOTE: This code has been modified from AWS Sagemaker Python:
# https://github.com/aws/sagemaker-python-sdk/blob/master/src/sagemaker/amazon/object2vec.py
#' @include r_utils.R
#' @import R6
#' @import sagemaker.common
#' @import sagemaker.mlcore
#' @import lgr
#' @title A general-purpose neural embedding algorithm that is highly customizable.
#' @description It can learn low-dimensional dense embeddings of high-dimensional objects. The embeddings
#' are learned in a way that preserves the semantics of the relationship between pairs of
#' objects in the original space in the embedding space.
#' @export
Object2Vec = R6Class("Object2Vec",
inherit = sagemaker.mlcore::AmazonAlgorithmEstimatorBase,
public = list(
#' @field repo_name
#' sagemaker repo name for framework
repo_name = "object2vec",
#' @field repo_version
#' version of framework
repo_version = 1,
#' @field MINI_BATCH_SIZE
#' The size of each mini-batch to use when training.
MINI_BATCH_SIZE = 32,
#' @field .module
#' mimic python module
.module = "sagemaker.amazon.object2vec",
#' @description Object2Vec is :class:`Estimator` used for anomaly detection.
#' This Estimator may be fit via calls to
#' :meth:`~sagemaker.amazon.amazon_estimator.AmazonAlgorithmEstimatorBase.fit`.
#' There is an utility
#' :meth:`~sagemaker.amazon.amazon_estimator.AmazonAlgorithmEstimatorBase.record_set`
#' that can be used to upload data to S3 and creates
#' :class:`~sagemaker.amazon.amazon_estimator.RecordSet` to be passed to
#' the `fit` call.
#' After this Estimator is fit, model data is stored in S3. The model
#' may be deployed to an Amazon SageMaker Endpoint by invoking
#' :meth:`~sagemaker.amazon.estimator.EstimatorBase.deploy`. As well as
#' deploying an Endpoint, deploy returns a
#' :class:`~sagemaker.amazon.Predictor` object that can be used for
#' inference calls using the trained model hosted in the SageMaker
#' Endpoint.
#' Object2Vec Estimators can be configured by setting hyperparameters.
#' The available hyperparameters for Object2Vec are documented below.
#' For further information on the AWS Object2Vec algorithm, please
#' consult AWS technical documentation:
#' https://docs.aws.amazon.com/sagemaker/latest/dg/object2vec.html
#' @param 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 accessing AWS resource.
#' @param instance_count (int): Number of Amazon EC2 instances to use
#' for training.
#' @param instance_type (str): Type of EC2 instance to use for training,
#' for example, 'ml.c4.xlarge'.
#' @param epochs (int): Total number of epochs for SGD training
#' @param enc0_max_seq_len (int): Maximum sequence length
#' @param enc0_vocab_size (int): Vocabulary size of tokens
#' @param enc_dim (int): Optional. Dimension of the output of the embedding
#' layer
#' @param mini_batch_size (int): Optional. mini batch size for SGD training
#' @param early_stopping_patience (int): Optional. The allowed number of
#' consecutive epochs without improvement before early stopping is
#' applied
#' @param early_stopping_tolerance (float): Optional. The value used to
#' determine whether the algorithm has made improvement between two
#' consecutive epochs for early stopping
#' @param dropout (float): Optional. Dropout probability on network layers
#' @param weight_decay (float): Optional. Weight decay parameter during
#' optimization
#' @param bucket_width (int): Optional. The allowed difference between data
#' sequence length when bucketing is enabled
#' @param num_classes (int): Optional. Number of classes for classification
#' @param training (ignored for regression problems)
#' @param mlp_layers (int): Optional. Number of MLP layers in the network
#' @param mlp_dim (int): Optional. Dimension of the output of MLP layer
#' @param mlp_activation (str): Optional. Type of activation function for the
#' MLP layer
#' @param output_layer (str): Optional. Type of output layer
#' @param optimizer (str): Optional. Type of optimizer for training
#' @param learning_rate (float): Optional. Learning rate for SGD training
#' @param negative_sampling_rate (int): Optional. Negative sampling rate
#' @param comparator_list (str): Optional. Customization of comparator
#' operator
#' @param tied_token_embedding_weight (bool): Optional. Tying of token
#' embedding layer weight
#' @param token_embedding_storage_type (str): Optional. Type of token
#' embedding storage
#' @param enc0_network (str): Optional. Network model of encoder "enc0"
#' @param enc1_network (str): Optional. Network model of encoder "enc1"
#' @param enc0_cnn_filter_width (int): Optional. CNN filter width
#' @param enc1_cnn_filter_width (int): Optional. CNN filter width
#' @param enc1_max_seq_len (int): Optional. Maximum sequence length
#' @param enc0_token_embedding_dim (int): Optional. Output dimension of token
#' embedding layer
#' @param enc1_token_embedding_dim (int): Optional. Output dimension of token
#' embedding layer
#' @param enc1_vocab_size (int): Optional. Vocabulary size of tokens
#' @param enc0_layers (int): Optional. Number of layers in encoder
#' @param enc1_layers (int): Optional. Number of layers in encoder
#' @param enc0_freeze_pretrained_embedding (bool): Optional. Freeze pretrained
#' embedding weights
#' @param enc1_freeze_pretrained_embedding (bool): Optional. Freeze pretrained
#' embedding weights
#' @param ... : base class keyword argument values.
initialize = function(role,
instance_count,
instance_type,
epochs,
enc0_max_seq_len,
enc0_vocab_size,
enc_dim=NULL,
mini_batch_size=NULL,
early_stopping_patience=NULL,
early_stopping_tolerance=NULL,
dropout=NULL,
weight_decay=NULL,
bucket_width=NULL,
num_classes=NULL,
mlp_layers=NULL,
mlp_dim=NULL,
mlp_activation=NULL,
output_layer=NULL,
optimizer=NULL,
learning_rate=NULL,
negative_sampling_rate=NULL,
comparator_list=NULL,
tied_token_embedding_weight=NULL,
token_embedding_storage_type=NULL,
enc0_network=NULL,
enc1_network=NULL,
enc0_cnn_filter_width=NULL,
enc1_cnn_filter_width=NULL,
enc1_max_seq_len=NULL,
enc0_token_embedding_dim=NULL,
enc1_token_embedding_dim=NULL,
enc1_vocab_size=NULL,
enc0_layers=NULL,
enc1_layers=NULL,
enc0_freeze_pretrained_embedding=NULL,
enc1_freeze_pretrained_embedding=NULL,
...){
private$.enc_dim = Hyperparameter$new("enc_dim", list(Validation$new()$ge(4), Validation$new()$le(10000)), "An integer in [4, 10000]", DataTypes$new()$int, obj = self)
private$.mini_batch_size = Hyperparameter$new("mini_batch_size", list(Validation$new()$ge(1), Validation$new()$le(10000)), "An integer in [1, 10000]", DataTypes$new()$int, obj = self)
private$.epochs = Hyperparameter$new("epochs", list(Validation$new()$ge(1), Validation$new()$le(100)), "An integer in [1, 100]", DataTypes$new()$int, obj = self)
private$.early_stopping_patience = Hyperparameter$new(
"early_stopping_patience", list(Validation$new()$ge(1), Validation$new()$le(5)), "An integer in [1, 5]", DataTypes$new()$int, obj = self
)
private$.early_stopping_tolerance = Hyperparameter$new(
"early_stopping_tolerance", list(Validation$new()$ge(1e-06), Validation$new()$le(0.1)), "A float in [1e-06, 0.1]", DataTypes$new()$float, obj = self
)
private$.dropout = Hyperparameter$new("dropout", list(Validation$new()$ge(0.0), Validation$new()$le(1.0)), "A float in [0.0, 1.0]", DataTypes$new()$float, obj = self)
private$.weight_decay = Hyperparameter$new("weight_decay", list(Validation$new()$ge(0.0), Validation$new()$le(10000.0)), "A float in [0.0, 10000.0]", DataTypes$new()$float, obj = self)
private$.bucket_width = Hyperparameter$new("bucket_width", list(Validation$new()$ge(0), Validation$new()$le(100)), "An integer in [0, 100]", DataTypes$new()$int, obj = self)
private$.num_classes = Hyperparameter$new("num_classes", list(Validation$new()$ge(2), Validation$new()$le(30)), "An integer in [2, 30]", DataTypes$new()$int, obj = self)
private$.mlp_layers = Hyperparameter$new("mlp_layers", list(Validation$new()$ge(1), Validation$new()$le(10)), "An integer in [1, 10]", DataTypes$new()$int, obj = self)
private$.mlp_dim = Hyperparameter$new("mlp_dim", list(Validation$new()$ge(2), Validation$new()$le(10000)), "An integer in [2, 10000]", DataTypes$new()$int, obj = self)
private$.mlp_activation = Hyperparameter$new(
"mlp_activation", Validation$new()$isin(c("tanh", "relu", "linear")), 'One of "tanh", "relu", "linear"', DataTypes$new()$str, obj = self
)
private$.output_layer = Hyperparameter$new(
"output_layer",
Validation$new()$isin(c("softmax", "mean_squared_error")),
'One of "softmax", "mean_squared_error"',
DataTypes$new()$str,
obj = self
)
private$.optimizer = Hyperparameter$new(
"optimizer",
Validation$new()$isin(c("adagrad", "adam", "rmsprop", "sgd", "adadelta")),
'One of "adagrad", "adam", "rmsprop", "sgd", "adadelta"',
DataTypes$new()$str,
obj = self
)
private$.learning_rate = Hyperparameter$new("learning_rate", list(Validation$new()$ge(1e-06), Validation$new()$le(1.0)), "A float in [1e-06, 1.0]", DataTypes$new()$float, obj = self)
private$.negative_sampling_rate = Hyperparameter$new(
"negative_sampling_rate", list(Validation$new()$ge(0), Validation$new()$le(100)), "An integer in [0, 100]", DataTypes$new()$int, obj = self
)
private$.comparator_list = Hyperparameter$new(
"comparator_list",
private$.list_check_subset(c("hadamard", "concat", "abs_diff")),
'Comma-separated of hadamard, concat, abs_diff. E.g. "hadamard,abs_diff"',
DataTypes$new()$str,
obj = self
)
private$.tied_token_embedding_weight = Hyperparameter$new(
"tied_token_embedding_weight", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
private$.token_embedding_storage_type = Hyperparameter$new(
"token_embedding_storage_type",
Validation$new()$isin(c("dense", "row_sparse")),
'One of "dense", "row_sparse"',
DataTypes$new()$str,
obj = self
)
private$.enc0_network = Hyperparameter$new(
"enc0_network",
Validation$new()$isin(c("hcnn", "bilstm", "pooled_embedding")),
'One of "hcnn", "bilstm", "pooled_embedding"',
DataTypes$new()$str,
obj = self
)
private$.enc1_network = Hyperparameter$new(
"enc1_network",
Validation$new()$isin(c("hcnn", "bilstm", "pooled_embedding", "enc0")),
'One of "hcnn", "bilstm", "pooled_embedding", "enc0"',
DataTypes$new()$str,
obj = self
)
private$.enc0_cnn_filter_width = Hyperparameter$new("enc0_cnn_filter_width", list(Validation$new()$ge(1), Validation$new()$le(9)), "An integer in [1, 9]", DataTypes$new()$int, obj = self)
private$.enc1_cnn_filter_width = Hyperparameter$new("enc1_cnn_filter_width", list(Validation$new()$ge(1), Validation$new()$le(9)), "An integer in [1, 9]", DataTypes$new()$int, obj = self)
private$.enc0_max_seq_len = Hyperparameter$new("enc0_max_seq_len", list(Validation$new()$ge(1), Validation$new()$le(5000)), "An integer in [1, 5000]", DataTypes$new()$int, obj = self)
private$.enc1_max_seq_len = Hyperparameter$new("enc1_max_seq_len", list(Validation$new()$ge(1), Validation$new()$le(5000)), "An integer in [1, 5000]", DataTypes$new()$int, obj = self)
private$.enc0_token_embedding_dim = Hyperparameter$new(
"enc0_token_embedding_dim", list(Validation$new()$ge(2), Validation$new()$le(1000)), "An integer in [2, 1000]", DataTypes$new()$int, obj = self
)
private$.enc1_token_embedding_dim = Hyperparameter$new(
"enc1_token_embedding_dim", list(Validation$new()$ge(2), Validation$new()$le(1000)), "An integer in [2, 1000]", DataTypes$new()$int, obj = self
)
private$.enc0_vocab_size = Hyperparameter$new("enc0_vocab_size", list(Validation$new()$ge(2), Validation$new()$le(3000000)), "An integer in [2, 3000000]", DataTypes$new()$int, obj = self)
private$.enc1_vocab_size = Hyperparameter$new("enc1_vocab_size", list(Validation$new()$ge(2), Validation$new()$le(3000000)), "An integer in [2, 3000000]", DataTypes$new()$int, obj = self)
private$.enc0_layers = Hyperparameter$new("enc0_layers", list(Validation$new()$ge(1), Validation$new()$le(4)), "An integer in [1, 4]", DataTypes$new()$int, obj = self)
private$.enc1_layers = Hyperparameter$new("enc1_layers", list(Validation$new()$ge(1), Validation$new()$le(4)), "An integer in [1, 4]", DataTypes$new()$int, obj = self)
private$.enc0_freeze_pretrained_embedding = Hyperparameter$new(
"enc0_freeze_pretrained_embedding", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
private$.enc1_freeze_pretrained_embedding = Hyperparameter$new(
"enc1_freeze_pretrained_embedding", list(), "Either True or False", DataTypes$new()$bool, obj = self
)
super$initialize(role, instance_count, instance_type, ...)
self$enc_dim = enc_dim
self$mini_batch_size = mini_batch_size
self$epochs = epochs
self$early_stopping_patience = early_stopping_patience
self$early_stopping_tolerance = early_stopping_tolerance
self$dropout = dropout
self$weight_decay = weight_decay
self$bucket_width = bucket_width
self$num_classes = num_classes
self$mlp_layers = mlp_layers
self$mlp_dim = mlp_dim
self$mlp_activation = mlp_activation
self$output_layer = output_layer
self$optimizer = optimizer
self$learning_rate = learning_rate
self$negative_sampling_rate = negative_sampling_rate
self$comparator_list = comparator_list
self$tied_token_embedding_weight = tied_token_embedding_weight
self$token_embedding_storage_type = token_embedding_storage_type
self$enc0_network = enc0_network
self$enc1_network = enc1_network
self$enc0_cnn_filter_width = enc0_cnn_filter_width
self$enc1_cnn_filter_width = enc1_cnn_filter_width
self$enc0_max_seq_len = enc0_max_seq_len
self$enc1_max_seq_len = enc1_max_seq_len
self$enc0_token_embedding_dim = enc0_token_embedding_dim
self$enc1_token_embedding_dim = enc1_token_embedding_dim
self$enc0_vocab_size = enc0_vocab_size
self$enc1_vocab_size = enc1_vocab_size
self$enc0_layers = enc0_layers
self$enc1_layers = enc1_layers
self$enc0_freeze_pretrained_embedding = enc0_freeze_pretrained_embedding
self$enc1_freeze_pretrained_embedding = enc1_freeze_pretrained_embedding
},
#' @description Return a :class:`~sagemaker.amazon.Object2VecModel` referencing the
#' latest s3 model data produced by this Estimator.
#' @param 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.
#' @param ... : Additional kwargs passed to the Object2VecModel constructor.
create_model = function(vpc_config_override="VPC_CONFIG_DEFAULT",
...){
return(Object2VecModel$new(
self$model_data,
self$role,
sagemaker_session=self$sagemaker_session,
vpc_config=self$get_vpc_config(vpc_config_override),
...
)
)
},
#' @description Set hyperparameters needed for training. This method will also
#' validate ``source_dir``.
#' @param records (RecordSet) – The records to train this Estimator on.
#' @param mini_batch_size (int or None) – The size of each mini-batch to use
#' when training. If None, a default value will be used.
#' @param job_name (str): Name of the training job to be created. If not
#' specified, one is generated, using the base name given to the
#' constructor if applicable.
.prepare_for_training = function(records,
mini_batch_size=NULL,
job_name=NULL){
if (is.null(mini_batch_size))
mini_batch_size = self$MINI_BATCH_SIZE
super$.prepare_for_training(
records, mini_batch_size=mini_batch_size, job_name=job_name
)
}
),
private = list(
# --------- User Active binding to mimic Python's Descriptor Class ---------
.epochs=NULL,
.enc_dim=NULL,
.mini_batch_size=NULL,
.early_stopping_patience=NULL,
.early_stopping_tolerance=NULL,
.dropout=NULL,
.weight_decay=NULL,
.bucket_width=NULL,
.num_classes=NULL,
.mlp_layers=NULL,
.mlp_dim=NULL,
.mlp_activation=NULL,
.output_layer=NULL,
.optimizer=NULL,
.learning_rate=NULL,
.negative_sampling_rate=NULL,
.comparator_list=NULL,
.tied_token_embedding_weight=NULL,
.token_embedding_storage_type=NULL,
.enc0_network=NULL,
.enc1_network=NULL,
.enc0_cnn_filter_width=NULL,
.enc1_cnn_filter_width=NULL,
.enc0_max_seq_len=NULL,
.enc1_max_seq_len=NULL,
.enc0_token_embedding_dim=NULL,
.enc1_token_embedding_dim=NULL,
.enc0_vocab_size=NULL,
.enc1_vocab_size=NULL,
.enc0_layers=NULL,
.enc1_layers=NULL,
.enc0_freeze_pretrained_embedding=NULL,
.enc1_freeze_pretrained_embedding=NULL,
# ---------
.list_check_subset = function(valid_super_list){
valid_superset = unique(valid_super_list)
validate = function(value){
if(!inherits(value, "character"))
return(FALSE)
val_list = lapply(split_str(value), trimws)
return(all(val_list %in% valid_superset))
}
return(validate)
}
),
active = list(
# --------- User Active binding to mimic Python's Descriptor Class ---------
#' @field epochs
#' Total number of epochs for SGD training
epochs = function(value){
if(missing(value))
return(private$.epochs$descriptor)
private$.epochs$descriptor = value
},
#' @field enc_dim
#' Dimension of the output of the embedding layer
enc_dim = function(value){
if(missing(value))
return(private$.enc_dim$descriptor)
private$.enc_dim$descriptor = value
},
#' @field mini_batch_size
#' mini batch size for SGD training
mini_batch_size = function(value){
if(missing(value))
return(private$.mini_batch_size$descriptor)
private$.mini_batch_size$descriptor = value
},
#' @field early_stopping_patience
#' The allowed number of consecutive epochs without
#' improvement before early stopping is applied
early_stopping_patience = function(value){
if(missing(value))
return(private$.early_stopping_patience$descriptor)
private$.early_stopping_patience$descriptor = value
},
#' @field early_stopping_tolerance
#' The value used to determine whether the algorithm
#' has made improvement between two consecutive epochs for early stopping
early_stopping_tolerance = function(value){
if(missing(value))
return(private$.early_stopping_tolerance$descriptor)
private$.early_stopping_tolerance$descriptor = value
},
#' @field dropout
#' Dropout probability on network layers
dropout = function(value){
if(missing(value))
return(private$.dropout$descriptor)
private$.dropout$descriptor = value
},
#' @field weight_decay
#' Weight decay parameter during optimization
weight_decay = function(value){
if(missing(value))
return(private$.weight_decay$descriptor)
private$.weight_decay$descriptor = value
},
#' @field bucket_width
#' The allowed difference between data sequence length when bucketing is enabled
bucket_width = function(value){
if(missing(value))
return(private$.bucket_width$descriptor)
private$.bucket_width$descriptor = value
},
#' @field num_classes
#' Number of classes for classification
num_classes = function(value){
if(missing(value))
return(private$.num_classes$descriptor)
private$.num_classes$descriptor = value
},
#' @field mlp_layers
#' Number of MLP layers in the network
mlp_layers = function(value){
if(missing(value))
return(private$.mlp_layers$descriptor)
private$.mlp_layers$descriptor = value
},
#' @field mlp_dim
#' Dimension of the output of MLP layer
mlp_dim = function(value){
if(missing(value))
return(private$.mlp_dim$descriptor)
private$.mlp_dim$descriptor = value
},
#' @field mlp_activation
#' Type of activation function for the MLP layer
mlp_activation = function(value){
if(missing(value))
return(private$.mlp_activation$descriptor)
private$.mlp_activation$descriptor = value
},
#' @field output_layer
#' Type of output layer
output_layer = function(value){
if(missing(value))
return(private$.output_layer$descriptor)
private$.output_layer$descriptor = value
},
#' @field optimizer
#' Type of optimizer for training
optimizer = function(value){
if(missing(value))
return(private$.optimizer$descriptor)
private$.optimizer$descriptor = value
},
#' @field learning_rate
#' Learning rate for SGD training
learning_rate = function(value){
if(missing(value))
return(private$.learning_rate$descriptor)
private$.learning_rate$descriptor = value
},
#' @field negative_sampling_rate
#' Negative sampling rate
negative_sampling_rate = function(value){
if(missing(value))
return(private$.negative_sampling_rate$descriptor)
private$.negative_sampling_rate$descriptor = value
},
#' @field comparator_list
#' Customization of comparator operator
comparator_list = function(value){
if(missing(value))
return(private$.comparator_list$descriptor)
private$.comparator_list$descriptor = value
},
#' @field tied_token_embedding_weight
#' Tying of token embedding layer weight
tied_token_embedding_weight = function(value){
if(missing(value))
return(private$.tied_token_embedding_weight$descriptor)
private$.tied_token_embedding_weight$descriptor = value
},
#' @field token_embedding_storage_type
#' Type of token embedding storage
token_embedding_storage_type = function(value){
if(missing(value))
return(private$.token_embedding_storage_type$descriptor)
private$.token_embedding_storage_type$descriptor = value
},
#' @field enc0_network
#' Network model of encoder "enc0"
enc0_network = function(value){
if(missing(value))
return(private$.enc0_network$descriptor)
private$.enc0_network$descriptor = value
},
#' @field enc1_network
#' Network model of encoder "enc1"
enc1_network = function(value){
if(missing(value))
return(private$.enc1_network$descriptor)
private$.enc1_network$descriptor = value
},
#' @field enc0_cnn_filter_width
#' CNN filter width
enc0_cnn_filter_width = function(value){
if(missing(value))
return(private$.enc0_cnn_filter_width$descriptor)
private$.enc0_cnn_filter_width$descriptor = value
},
#' @field enc1_cnn_filter_width
#' CNN filter width
enc1_cnn_filter_width = function(value){
if(missing(value))
return(private$.enc1_cnn_filter_width$descriptor)
private$.enc1_cnn_filter_width$descriptor = value
},
#' @field enc0_max_seq_len
#' Maximum sequence length
enc0_max_seq_len = function(value){
if(missing(value))
return(private$.enc0_max_seq_len$descriptor)
private$.enc0_max_seq_len$descriptor = value
},
#' @field enc1_max_seq_len
#' Maximum sequence length
enc1_max_seq_len = function(value){
if(missing(value))
return(private$.enc1_max_seq_len$descriptor)
private$.enc1_max_seq_len$descriptor = value
},
#' @field enc0_token_embedding_dim
#' Output dimension of token embedding layer
enc0_token_embedding_dim = function(value){
if(missing(value))
return(private$.enc0_token_embedding_dim$descriptor)
private$.enc0_token_embedding_dim$descriptor = value
},
#' @field enc1_token_embedding_dim
#' Output dimension of token embedding layer
enc1_token_embedding_dim = function(value){
if(missing(value))
return(private$.enc1_token_embedding_dim$descriptor)
private$.enc1_token_embedding_dim$descriptor = value
},
#' @field enc0_vocab_size
#' Vocabulary size of tokens
enc0_vocab_size = function(value){
if(missing(value))
return(private$.enc0_vocab_size$descriptor)
private$.enc0_vocab_size$descriptor = value
},
#' @field enc1_vocab_size
#' Vocabulary size of tokens
enc1_vocab_size = function(value){
if(missing(value))
return(private$.enc1_vocab_size$descriptor)
private$.enc1_vocab_size$descriptor = value
},
#' @field enc0_layers
#' Number of layers in encoder
enc0_layers = function(value){
if(missing(value))
return(private$.enc0_layers$descriptor)
private$.enc0_layers$descriptor = value
},
#' @field enc1_layers
#' Number of layers in encoder
enc1_layers = function(value){
if(missing(value))
return(private$.enc1_layers$descriptor)
private$.enc1_layers$descriptor = value
},
#' @field enc0_freeze_pretrained_embedding
#' Freeze pretrained embedding weights
enc0_freeze_pretrained_embedding = function(value){
if(missing(value))
return(private$.enc0_freeze_pretrained_embedding$descriptor)
private$.enc0_freeze_pretrained_embedding$descriptor = value
},
#' @field enc1_freeze_pretrained_embedding
#' Freeze pretrained embedding weights
enc1_freeze_pretrained_embedding = function(value){
if(missing(value))
return(private$.enc1_freeze_pretrained_embedding$descriptor)
private$.enc1_freeze_pretrained_embedding$descriptor = value
}
),
lock_objects = F
)
#' @title Reference Object2Vec s3 model data.
#' @description Calling :meth:`~sagemaker.model.Model.deploy` creates an Endpoint and returns a
#' Predictor that calculates anomaly scores for datapoints.
#' @export
Object2VecModel = R6Class("Object2VecModel",
inherit = sagemaker.mlcore::Model,
public = list(
#' @description Initialize Object2VecModel class
#' @param model_data (str): The S3 location of a SageMaker model data
#' ``.tar.gz`` file.
#' @param role (str): An AWS IAM role (either name or full ARN). The Amazon
#' SageMaker training jobs and APIs that create Amazon SageMaker
#' endpoints use this role to access training data and model
#' artifacts. After the endpoint is created, the inference code
#' might use the IAM role, if it needs to access an AWS resource.
#' @param sagemaker_session (sagemaker.session.Session): Session object which
#' manages interactions with Amazon SageMaker APIs and any other
#' AWS services needed. If not specified, the estimator creates one
#' using the default AWS configuration chain.
#' @param ... : Keyword arguments passed to the ``FrameworkModel``
#' initializer.
initialize = function(model_data,
role,
sagemaker_session=NULL,
...){
sagemaker_session = sagemaker_session %||% sagemaker.core::Session$new()
image_uri = sagemaker.core::ImageUris$new()$retrieve(
Object2Vec$public_fields$repo_name,
sagemaker_session$paws_region_name,
version=Object2Vec$public_fields$repo_version
)
super$initialize(
image_uri,
model_data,
role,
predictor_cls=Predictor,
sagemaker_session=sagemaker_session,
...
)
}
),
lock_objects = F
)
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.