sageworks
sageworks copied to clipboard
SuperCowPowers
SageWorks: An easy to use Python API for creating and deploying AWS SageMaker Models
Right now the tables for the Dashboard, DataSources, FeatureSets, Models, Endpoint are sorted off of name (I think) will be more useful to have them sorted via Created Date. Also...
Might be nice to include an event bridge viewer in the web components. Perhaps have it's own SQS.
A known issue with the SageMaker Python API https://github.com/aws/sagemaker-python-sdk/issues/2916
We should have this be part of our 'SageWorks' AWS Deployment using Python CDK. Note: This is future, maybe stuff... it's kinda a lets look at this and see how...
Have Documentation about how the SageWorks Visibility Scope is used. SageWorks has a 'visibility scope' that gets set... if your data is 'here' and models are 'there' you simply add...
Building complicated processing pipelines will be WAY easier with Lambdas and StepFunctions. Do a couple of kewl 'demo' Step Functions.
https://github.com/scikit-learn-contrib/skdag
One of the interesting things about the AWS Feature Store is that the data is Parquet partitioned by `EventTime`. We should construct a kewl demo of how that might be...
We're mostly thinking/using features sets in offline/batch mode, so lets think of some kewl use case where we use it in online mode: - Compound Lookup - Features - Tags...
https://dash.plotly.com/cytoscape/layout See 'Physics Base Layout' toward the bottom of the page. That that example, create a Python Class and replace the circles with the SageWorks icon set for data source,...
Metadata
Owner
Metadata
SageWorks: An easy to use Python API for creating and deploying AWS SageMaker Models