Class: Aws::SageMaker::Types::AnnotationConsolidationConfig

Inherits:
Struct
  • Object
show all
Includes:
Aws::Structure
Defined in:
lib/aws-sdk-sagemaker/types.rb

Overview

Note:

When making an API call, you may pass AnnotationConsolidationConfig data as a hash:

{
  annotation_consolidation_lambda_arn: "LambdaFunctionArn", # required
}

Configures how labels are consolidated across human workers.

See Also:

Instance Attribute Summary collapse

Instance Attribute Details

#annotation_consolidation_lambda_arn ⇒ String

The Amazon Resource Name (ARN) of a Lambda function implements the logic for annotation consolidation.

For the built-in bounding box, image classification, semantic segmentation, and text classification task types, Amazon SageMaker Ground Truth provides the following Lambda functions:

  • *Bounding box* - Finds the most similar boxes from different workers based on the Jaccard index of the boxes.

    ‘arn:aws:lambda:us-east-1:432418664414:function:ACS-BoundingBox`

    ‘arn:aws:lambda:us-east-2:266458841044:function:ACS-BoundingBox`

    ‘arn:aws:lambda:us-west-2:081040173940:function:ACS-BoundingBox`

    ‘arn:aws:lambda:eu-west-1:568282634449:function:ACS-BoundingBox`

    ‘arn:aws:lambda:ap-northeast-1:477331159723:function:ACS-BoundingBox`

  • *Image classification* - Uses a variant of the Expectation Maximization approach to estimate the true class of an image based on annotations from individual workers.

    ‘arn:aws:lambda:us-east-1:432418664414:function:ACS-ImageMultiClass`

    ‘arn:aws:lambda:us-east-2:266458841044:function:ACS-ImageMultiClass`

    ‘arn:aws:lambda:us-west-2:081040173940:function:ACS-ImageMultiClass`

    ‘arn:aws:lambda:eu-west-1:568282634449:function:ACS-ImageMultiClass`

    ‘arn:aws:lambda:ap-northeast-1:477331159723:function:ACS-ImageMultiClass`

  • *Semantic segmentation* - Treats each pixel in an image as a multi-class classification and treats pixel annotations from workers as “votes” for the correct label.

    ‘arn:aws:lambda:us-east-1:432418664414:function:ACS-SemanticSegmentation`

    ‘arn:aws:lambda:us-east-2:266458841044:function:ACS-SemanticSegmentation`

    ‘arn:aws:lambda:us-west-2:081040173940:function:ACS-SemanticSegmentation`

    ‘arn:aws:lambda:eu-west-1:568282634449:function:ACS-SemanticSegmentation`

    ‘arn:aws:lambda:ap-northeast-1:477331159723:function:ACS-SemanticSegmentation`

  • *Text classification* - Uses a variant of the Expectation Maximization approach to estimate the true class of text based on annotations from individual workers.

    ‘arn:aws:lambda:us-east-1:432418664414:function:ACS-TextMultiClass`

    ‘arn:aws:lambda:us-east-2:266458841044:function:ACS-TextMultiClass`

    ‘arn:aws:lambda:us-west-2:081040173940:function:ACS-TextMultiClass`

    ‘arn:aws:lambda:eu-west-1:568282634449:function:ACS-TextMultiClass`

    ‘arn:aws:lambda:ap-northeast-1:477331159723:function:ACS-TextMultiClass`

For more information, see [Annotation Consolidation].

[1]: docs.aws.amazon.com/sagemaker/latest/dg/sms-annotation-consolidation.html

Returns:

  • (String) 


511
512
513
514
 
# File 'lib/aws-sdk-sagemaker/types.rb', line 511

class AnnotationConsolidationConfig < Struct.new(
  :annotation_consolidation_lambda_arn)
  include Aws::Structure
end