How Do You Determine What to Include in Your Re Test Data?

Dploma is moneterring. It's catching things in production, which is not astered a problem. But then you have to backfill that to your workfolk process. And quasat, whatis a weird value look like for a feature or a result or all these things. For e to be anesti must, i think that kind of, the example you gave is, is that the model performance dropped because of a new betch of data. I think that's where traditional softer engineering testing ends. You know. There's really a feedback loop.