Databricks Machine Learning Associate Question 115
Single answerYou are using MLflow Tracking in Databricks and need to locate the code that was executed for a specific run recorded in MLflow. How can you find this information in the MLflow UI?
- A
Navigate to the 'Artifacts' tab of the run and look for a source code snapshot or script file.
- B
Inspect the 'Tags' section of the run for a tag named 'mlflow.source.code'.
- C
Open the 'Source' tab of the run to view the code.
- D
Check the 'Parameters' section of the run for the code used to generate the run.
Show answer and explanation
Correct answer: A
Explanation
The 'Artifacts' tab in the MLflow UI is where logged files, including snapshots of source code or scripts, are stored during a run. If the code was logged as an artifact (e.g., using mlflow.log_artifact or through automatic logging features), it can be found here. This makes the 'Artifacts' tab the correct location to locate executed code.
- A. Correct.
Correct. The 'Artifacts' tab in the MLflow UI often contains a snapshot of the source code (e.g., a script file or notebook) that was used during the run, provided the code was logged during the run.
- B. Incorrect.
Incorrect. The 'Tags' section does not store the actual code but may contain metadata about the source, such as the origin of the run (e.g., Jupyter notebook or Databricks job).
- C. Incorrect.
Incorrect. MLflow does not have a dedicated 'Source' tab. The source information is typically found in tags or artifacts, not in a separate tab.
- D. Incorrect.
Incorrect. The 'Parameters' section stores input parameters for the run but does not contain the code that was executed.