Types of inference · Question 12 of 12
Which requirements favour asynchronous over real-time inference?
3 answers are correct. Hard
Show answer
Answer
- A Processing times of several minutes per request
- B The result may be collected later from Amazon S3
- D Inputs of several hundred megabytes
Explanation
Asynchronous inference is built for large payloads and long run times: requests go into a queue and the result is stored in S3, optionally with a notification. For answers within milliseconds or an interactive dialogue, real-time inference is the right choice.