Total points 3
1.
Question 1
Which of the following are advantages of batch inference?

1 / 1 point

> You can use complex machine learning models to improve accuracy since there is no constraint on inference time.

Correct
Nice job!


> You don’t need caching strategies so costs decrease.

Correct
Great work!


> You can wait for data retrieval to make predictions since they are not made available in real time.

Correct
Excellent!


You achieve shorter latency of predictions.

2.
Question 2
True or False: The most important metric to optimize while performing batch predictions is throughput.

1 / 1 point

False


> True

Correct
Yes! Triton Inference Server allows deployment of models from any framework, from local storage, Google Cloud Platform, or AWS S3.

3.
Question 3
How can you save inference costs when generating recommendations on e-commerce?

1 / 1 point

Generating them every time a user logs in.


> Generating them on a schedule.


Using hardware accelerators.

Correct
That's right!