A data scientist registers a Python XGBoost model in SAS Model Manager by uploading a ZIP file. After registration, she wants to run the model's score code on a test dataset inside SAS Model Manager. What must be true for the scoring to succeed?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because SAS Model Manager uses SAS Open Model Interface conventions for Python models: the score.py file must define a scoring function (typically named score) that accepts input data as a pandas DataFrame and returns a pandas DataFrame containing predicted values. SAS Model Manager calls this function during test scoring.
Full explanation below image
Full Explanation
B is correct because SAS Model Manager uses SAS Open Model Interface conventions for Python models: the score.py file must define a scoring function (typically named score) that accepts input data as a pandas DataFrame and returns a pandas DataFrame containing predicted values. SAS Model Manager calls this function during test scoring. A is partially true (requirements.txt is included) but the scoring convention requirement is the key technical point. C is wrong; PROC ASTORE is for SAS-native ASTORE objects, not Python pickle files. D is wrong; SAS Model Manager does not require a Jupyter kernel connection.