Skip to content

Release 0.8.29

Need Help?

The SuperCowPowers team is happy to give any assistance needed when setting up AWS and Workbench. So please contact us at workbench@supercowpowers.com or on chat us up on Discord

The Workbench framework continues to flex to support different real world use cases when operating a set of production machine learning pipelines.

Note: These release notes cover the changes from 0.8.27 to 0.8.29

General

Locking AWS Model Training Image: AWS will randomly update the images associated with training and model registration. In particular the SKLearn Estimator has been updated into a non-working state for our use cases. So for both training and registration we're now explicitly specifying the image that we want to use.

 self.estimator = SKLearn(
     ...
     framework_version="1.2-1",
     image_uri=image,  # New
 )

API Changes

  • delete() --> class.delete(uuid)

    We've changed the API for deleting artifacts in AWS (DataSource, FeatureSet, etc). This is part of our efforts to minimize race-conditions when objects are deleted. 

    my_model = Model("xyz")  # Creating object
my_model.delete()        # just to delete

<Now just one line>
Model.delete("xyz")      # Delete

Minor Improvements

Bulk Delete: Added a Bulk Delete utility

from workbench.utils.bulk_utils import bulk_delete

delete_list = [("DataSource", "abc"), ("FeatureSet", "abc_features")]
bulk_delete(delete_list)

Questions?

The SuperCowPowers team is happy to answer any questions you may have about AWS and Workbench. Please contact us at workbench@supercowpowers.com or on chat us up on Discord