Azure Data Scientists Associate Practice Exam

In training scripts, what is necessary to log the target metric "AUC" using MLflow?

1. Use a print() statement to display AUC

2. Use logging.info() to log the AUC

3. Use mlflow.log_metric() to log the AUC

4. Use an assert statement

The correct answer is: Use mlflow.log_metric() to log the AUC

In training scripts, to effectively log the target metric "AUC" using MLflow, utilizing the function that specifically handles the recording of metrics is essential. The function mlflow.log_metric() is designed for this purpose, allowing users to log the value of a metric at a specific step in their training process. By using this function, the AUC value will be accurately captured and stored in MLflow's tracking server, which enables easy monitoring and comparison of model performance over different training runs. This function offers a structured way to record not just the AUC but any other relevant metrics, ensuring that the data is organized and can be retrieved for analysis and visualization later. This capability is crucial when iterating on models, allowing data scientists to understand how adjustments to their algorithms impact performance metrics over time. The other options do not fit the purpose of logging the AUC in a way that integrates with MLflow's tracking capabilities. Using a print() statement would merely output the value to the console without storing it for future reference. Likewise, logging with logging.info() might capture the AUC value in standard logs but would not facilitate its retrieval in a structured manner through MLflow. An assert statement is used for assertions in code to check if a condition holds true