easy5 pts · first answer countsdata-validationpipelines

Bad data reached training

An upstream bug filled a column with zeros, and the weekly retrain shipped a worse model. What would have caught it earliest?

A.A larger modelB.Automated data validation checks (schema, ranges, null and zero rates) before trainingC.More GPUsD.Retraining daily instead of weekly

Discussion 0

Join free to post in the discussion.

Bad data reached training

Discussion 0

Related questions