Skip to content

SageMaker Canvas > Dataset Auto Update & Automatically Generate Prediction

0

Could you explain how the Auto Update feature in Canvas Datasets and the Automatic Generate Prediction feature in Canvas My Models work in terms of automation? Also, since these features are currently not working on my end, would it be possible for me to replicate the same functionality using EventBridge and a Lambda function?

asked a year ago134 views

1 Answer
0

In Amazon SageMaker Canvas, you can create automations that update your dataset and generate predictions on a schedule, reducing manual work.

For dataset auto-updates, you can configure Canvas to check for new files at a specified frequency. When you set up automatic updates, you specify an Amazon S3 location where Canvas will look for new files. Each update creates a new version of your dataset, with only the latest version being usable for model building or predictions. You can configure up to 20 automatic update configurations.

For automatic batch predictions, you can set up a workflow where predictions are generated whenever your dataset is updated. To configure this:

  1. Go to "My models" in Canvas
  2. Select your model
  3. Choose the "Predict" tab
  4. Select "Batch prediction" and choose "Automatic"
  5. Select the dataset you want to use for automated predictions

When you set this up, Canvas runs a batch prediction job immediately, and then automatically runs another job each time your dataset is updated (either manually or through auto-updates). This creates an automated workflow where new data leads to new predictions.

Important limitations to note:

  • These automations only run while you're logged into the Canvas application and pause when you log out
  • There may be a delay (approximately 10 minutes) before job history is updated in the interface
  • When a dataset is automatically updated, models do not automatically rebuild with the new data

Regarding replicating this functionality with EventBridge and Lambda: While not explicitly covered in the documentation provided, you could potentially create a custom solution using these AWS services to monitor for new data in S3, trigger dataset updates, and initiate batch prediction jobs through the SageMaker APIs. This would require custom code but could provide similar functionality without the Canvas application's limitations.
Sources
Configure automatic updates for a dataset - Amazon SageMaker
if sagemaker dataset gets a new version automatically, does the model automatically rebuild? | AWS re:Post
How to manage automations - Amazon SageMaker
Make automatic batch predictions - Amazon SageMaker AI
sagemaker canvas dataset auto update, when is job history updated? | AWS re:Post

answered a year ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.