How to enable automated data labeling for custom labeling job?

Question

Hi,  
I need to label large dataset which has around 200000 records and this doesn't fall into any of the prebuilt labeling job categories. So trying to create custom labeling job with automated data labeling enabled. But the configuration to enable the automated data labeling doesn't appear for custom labelling job. Any reference to proceed further asap would be of great help. Thank you.

Answer

Hi there. I'm an engineer at AWS.  
  
Automatic data labeling is only available for the first-party workflows: Image Classification, Text Classification, and Object Detection. As an alternative, you might consider orchestrating a series of labeling jobs with an intermediate step in which you train a machine learning model and perform inference to carry out an auto labeling step. We have a blog post that describes this "active learning" process to give you some ideas as to how you might accomplish this (https://aws.amazon.com/blogs/machine-learning/annotate-data-for-less-with-amazon-sagemaker-ground-truth-and-automated-data-labeling/).  
  
Thank you for being a valued customer and for bringing this feature request to our attention.

How to enable automated data labeling for custom labeling job?

関連するコンテンツ