Azure Data Scientists Associate Practice Exam

What functions must the scoring script for a real-time inferencing service include?

1. main() and score()

2. base() and train()

3. init() and run()

4. validate() and predict()

The correct answer is: init() and run()

The scoring script for a real-time inferencing service in Azure machine learning must include the init() and run() functions. The init() function initializes the model and is called once when the service starts. This is where you can load your trained model into memory, making it ready for inference requests. The run() function is called for each incoming data request, where it processes the input data and generates predictions based on the model loaded in the init() function. This separation allows for efficient handling of multiple requests without the need to reload the model repeatedly. This choice reflects the core requirements for setting up real-time inferencing within the Azure ecosystem, emphasizing the importance of these functions in ensuring smooth and effective operations in a deployed model.