A standard managed model deployment loads an artifact but cannot find its inference implementation. Which file should be reviewed?

1. Only the training dataset CSV
2. score.py
3. Only the README's title
4. Only a screenshot of notebook output