Databricks Machine Learning Professional Question 84
Single answerYou are managing a Databricks Machine Learning project and need to use MLflow model registry to track the lifecycle of your models. You have a model that is ready for production deployment but has not yet been tested in a staging environment. Which model stage should you assign to this model to indicate it is ready for testing in staging?
- A
None
- B
Staging
- C
Production
- D
Archived
Show answer and explanation
Correct answer: B
Explanation
In MLflow, the 'Staging' stage is specifically designed for models that are ready to be tested in a staging environment. This stage allows teams to validate the model before promoting it to 'Production' for live deployment. The other stages ('None', 'Production', and 'Archived') do not fit the scenario described.
- A. Incorrect.
The 'None' stage is the default stage when a model is first registered in the MLflow model registry. It does not indicate readiness for deployment or testing.
- B. Correct.
The 'Staging' stage is used to indicate that a model is ready for testing in a staging environment before it is deployed to production. This is the correct stage for your scenario.
- C. Incorrect.
The 'Production' stage is used for models that have already been tested and approved for deployment in a production environment. This does not fit the scenario where the model is not yet tested.
- D. Incorrect.
The 'Archived' stage is used for models that are no longer in active use. It is not relevant for a model that is ready for testing in staging.