sageworks
sageworks copied to clipboard
SageWorks: An easy to use Python API for creating and deploying AWS SageMaker Models
# Pull Request Checklist Thank you for contributing to the SuperCowPowers SageWorks project. Please make sure you've completed the following tasks before submitting your pull request: #### GENERAL - [x]...
We need to test out the API key functionality: - Expiration (Nice message when it expires) - Features Flags - Ensure AWS Account ID check
So the addition of kwargs is a nice bit of flexibility so that plugins can can in additional information, but looks like we still have a wrinkle with the implementation....
Let's add a `set_input` part to model onboarding, also if we're in a situation where the FeatureSet doesn't exist we need to think of a marker/flag that means the model...
Now that the model details markdown doesn't include model metrics (which is good), it can probably be more general. Right now it's a bit awkward when used in the Endpoint...
Right now the tag display on the Dashboard look 'not good'. So review how the dashboard is processing tags and make the display nicer. Tags: table:physchemproperty:processing:featurized:assay:solubility:feature_endpoint:smiles-to-rdkit-modred:date:2024_03_01
The 'query' link for the datasets seems to work fine, but the query link for FeatureSets all say 'unknown'.
Our test suite is giving deprecation warnings.. so run `tox` review the deprecation warnings and try to fix/address all of them.
After creating a DataSource with either a file or DataFrame input, have the ability to set an input 'path' (s3 whatever).