Practice English vocabulary for ML model versioning: MLflow registry, model stages, promoting models from staging to production, and champion/challenger patterns.
0 / 5 completed
1 / 5
Your team tracks every experiment and model artifact using MLflow. Which phrase is correct?
'The model is versioned in MLflow' means MLflow is used to track and store different versions of the model along with their parameters, metrics, and artifacts.
2 / 5
A central repository stores all trained model versions with their metadata and allows you to retrieve any previous version. What is this called?
A 'model registry' is a centralized store for managing the lifecycle of ML models — storing versions, tracking their status, and controlling promotion between stages.
3 / 5
In the model registry, a model can be marked as 'Staging', 'Production', or 'Archived'. What are these called?
'Model stages' in MLflow (and similar registries) represent the lifecycle state of a model version: Staging (testing), Production (serving live traffic), or Archived (retired).
4 / 5
After validation passes, the team moves the model from the testing environment to the live serving environment. Which phrase is correct?
'We promote the model from staging to production' is the standard phrasing. 'Promote' is the specific verb used for advancing a model version through registry stages.
5 / 5
The current best-performing model is serving production traffic. A new model is being tested against it with a small percentage of traffic to see if it performs better. What are these two models called?
In ML serving, the 'champion' is the current production model and the 'challenger' is a new model being tested against it. If the challenger outperforms the champion, it gets promoted.