awesome-mlops
awesome-mlops copied to clipboard
kelvins
add robust intelligence
What is this tool for?
At a high level, focused on "Proactively eliminate failure at every stage of the model lifecycle for ML integrity."
Specifically, does this in a few different ways, which is why I added to multiple categories. One way is by validating data (most similar to great expectations). Another is through model fairness and privacy, where it tests not only for generic bias/fairness but also checks to make sure compliant with several pieces of regulation. For testing and validating models, it has a suite of tests in many different categories. And it can do this testing across time as well, which puts in the same category as other monitoring tools (Visual Analysis and Debugging)
What's the difference between this tool and similar ones?
For pretty much every category the main difference is automatic profiling and configuration of the checks it is performing, allowing for faster onboarding and usage (you don't have to configure a million different monitors)
Meta note: I know the contribution guide say one link per PR, but given that this tool falls under multiple categories I figured most concrete to do in same PR? Happy to split up if that is preferred though!
Anyone who agrees with this pull request could submit an Approve review to it.
